Investigating Division Quest: A Text-Based Adventure Narrative

Georgia Southern University

Georgia Southern Commons

Electronic Theses and Dissertations Jack N. Averitt College of Graduate Studies

Spring 2024

Investigating Division Quest: A Text-Based Adventure

Narrative Game for Overcoming Di<iculties of Mitosis

Learning

Amanda G. Conner

Follow this and additional works at: https://digitalcommons.georgiasouthern.edu/etd

Part of the Cell Biology Commons, Educational Assessment, Evaluation, and Research

Commons, Educational Technology Commons, Online and Distance Education Commons,

Scholarship of Teaching and Learning Commons, and the Science and Mathematics

Education Commons

Recommended Citation

Conner, Amanda G., "Investigating Division Quest: A Text-Based Adventure Narrative Game

for Overcoming Di<iculties of Mitosis Learning" (2024).

Electronic Theses and

Dissertations

. 2781.

https://digitalcommons.georgiasouthern.edu/etd/2781

This thesis (open access) is brought to you for free and open access by the Jack N. Averitt College

of Graduate Studies at Georgia Southern Commons. It has been accepted for inclusion in Electronic

Theses and Dissertations by an authorized administrator of Georgia Southern Commons. For more

information, please contact [email protected].

INVESTIGATING DIVISION QUEST: A TEXT-BASED ADVENTURE NARRATIVE GAME

FOR OVERCOMING DIFFICULTIES OF MITOSIS LEARNING

AMANDA CONNER

(Under the Direction of Sue Ellen DeChenne-Peters)

ABSTRACT

Mitosis learning is a source of struggle for many introductory biology students. Much of this

difficulty stems from an overwhelming amount of terminology and moving pieces as well as the

abstract nature of cellular processes. Game-based learning (GBL) has been used in other contexts

to reduce cognitive load and provide relatable experiences upon which students can construct

their mental models. However, commercial GBL is often expensive and difficult to align with

individual instructor teaching preferences and learning objectives. In this study, we use a mixed

methods approach to explore the use of a GBL tool made in Twine to overcome some difficulties

of mitosis learning and reduce the barriers to using GBL. We conducted an autoethnography of

game development in conjunction with gathering pre/post test data from introductory biology

students to compare the posttest scores of students who played the game against those who

completed instructor activities, students who played the game in-class against those who played

it out-of-class, and students who fully completed the game against those who never played the

game. Results show that Division Quest promoted higher learning surrounding the visual

identification and nature of duplicated chromosomes (likely due to narrative and visual game

elements), but inhibited learning surrounding visual identification of mitotic phase from a

micrograph (likely due to misalignment of cell images used in the game and those used in

assessment). On all other measured criteria, students who played Division Quest learned

comparably to students who completed an in-class active learning activity or an at-home pre-lab

assignment. Our findings support the use of Twine GBL tools for mitosis learning and highlight

game features that appear to facilitate (or detract from) learning. Finally, we discuss future game

changes to address the game’s observed shortcomings and implications for the development of

similar GBL tools.

INDEX WORDS: Game-based learning, Mitosis learning, Twine, Games for learning, Narrative

games for learning, Cognitive load, Constructivism

INVESTIGATING DIVISION QUEST: A TEXT-BASED ADVENTURE NARRATIVE GAME

FOR OVERCOMING DIFFICULTIES OF MITOSIS LEARNING

AMANDA CONNER

B.S., Georgia Southern University, 2022

A Thesis Submitted to the Graduate Faculty of Georgia Southern University

in Partial Fulfillment of the Requirements for the Degree

MASTER OF SCIENCE

COLLEGE OF SCIENCE AND MATHEMATICS

AMANDA CONNER

INVESTIGATING DIVISION QUEST: A TEXT-BASED ADVENTURE NARRATIVE GAME

FOR OVERCOMING DIFFICULTIES OF MITOSIS LEARNING

AMANDA CONNER

Major Professor: Sue Ellen DeChenne-Peters

Committee: Robert Terry

Jennifer Zettler

Electronic Version Approved:

May 2024

DEDICATION

This thesis is dedicated to those whose unwavering love and support have been the mainstay of

my academic journey.

To my amazing husband, Daniel Conner: you have been and remain to be the most steadfast

foundation of my life. Your constant encouragement and praise have lifted me out of many a

deep rut that I’ve dug for myself. I thank you for every meal deposited silently on my desk at no

request, every backpack, bottle of water, and to-go breakfast taken to my car while I rushed

through my morning routine because I overslept, and all the little ways you remind me “you’re

doing great”.

To my outstanding online gaming friends: your camaraderie, companionship, and enthusiasm

provided much-needed perspective and motivation throughout this journey and fueled my

determination to see it through, even in the late nights and deepest doubts. Special thanks go to

Kara Scott, Mike Owen, Gio Contreras, Preacher Kotas, and Joe Cagle. I hope you all know how

much your support and friendship means to me.

To my beloved family, especially my mother and father, John and Kristie McGrath, and mother-

and father-in-law, Matt and Tami Conner: your unwavering love and support does not go

unnoticed. Your willingness and flexibility to step in and care for my daughter allowed me to

focus wholeheartedly on my studies while knowing that she is safe and loved in your capable

hands. Thank you for being pillars of my and your granddaughter’s support system.

Finally, to my exceptional daughter Laura: though you are too young right now, I hope that one

day you read this and know that becoming your mother has been the most transformative

experience of my entire life and was a cornerstone in the execution of this project. Your radiant

smile, boundless energy, and enthusiasm for life remind me of the beauty and joy in every

moment. You give me the motivation to embrace each day as an opportunity for growth and

strive to be the best version of myself. Being your mom is the greatest privilege of my life, and I

am endlessly grateful for the love and purpose you bring into my world.

This thesis stands as a testament to the love, encouragement, and sacrifices of each of you, who

have shaped and enriched my life in profound ways.

With love and gratitude,

Amanda

ACKNOWLEDGMENTS

Completing a project as interdisciplinary as this one is a monumental task that would not have

been possible without a team of dedicated individuals. First, I would like to extend my endless

gratitude to the members of my thesis committee. Each of them have invested their most

valuable resources – time and energy – in me and my studies from the early stages of my

academic career, and I know that I can continue to depend on their support as my career

progresses. Dr. Sue Ellen DeChenne-Peters has been my safe harbor and core of my academic

support network; I would not have been able to persist through the challenges of graduate school

without her care and support in my research, teaching, coursework, and personal life. Dr. Robert

Terry introduced me to game development in Twine and fostered my growth as a game

developer and researcher, as well as shared with me his passion for game-based learning,

innovative pedagogical tools, and ensuring accessibility in education. Likewise, Dr. Jennifer

Zettler has invested in my growth as a research scientist, both in content biology and biology

education, and brought a unique perspective on my work that enabled me to form deeper

connections between the disparate disciplines of my project and informed ways to present my

findings that will be accessible to practitioners outside of game-based learning and discipline-

based education research.

Next, I cannot thank the instructors and students who participated in our data collection enough.

Georgia Southern University, and the Department of Biology in particular, have endured several

tumultuous years in the wake of consolidation with Armstrong State University. While still

navigating the complexities of our institution’s new challenges, multiple instructors went out of

their way to make space for my research in their classrooms and provide advice on how to adapt

data collection to the changing circumstances. Dr. Geneva DeMars was the first to agree to

participate and oversaw all data collected on the Armstrong campus, and she did both

enthusiastically. Similarly, Dr. Lace Svec took the time to modify her lab prep materials,

coordinate with other lab instructors to collect all data on the Statesboro campus, and orient me

to her course materials.

Last but certainly not least, Laura Griffiths, former lab coordinator, was instrumental in the

acquisition of micrographs for the game and offered frequent insight and support throughout the

project. Similarly, the undergraduate research students in our lab, Bodhi Martin, Richmond

Boadu, and Kendall Lane kept up with the progress of my research while conducting research of

their own and contributed much to the early stages of game development. Division Quest would

not be what it is without their insightful and detailed input.

TABLE OF CONTENTS

Page

DEDICATION ...………………………………………………………………………………….2

ACKNOWLEDGMENTS ..............................................................................................................4

LIST OF TABLES ..........................................................................................................................8

LIST OF FIGURES ......................................................................................................................10

CHAPTER

1 INTRODUCTION …...………………………………………………………………..11

MISCONCEPTIONS IN MITOSIS LEARNING …….………………………...11

GAMES FOR LEARNING …………………………...………………………...12

GAME-BASED LEARNING THEORETICAL FRAMEWORKS...…...………16

RESEARCH QUESTIONS …..……………………….…………………...……18

2 EFFICACY OF GBL-BASED APPROACH ...…………….……………….…...……19

METHODS ……..………………………….........................................................19

RESULTS ………...…………………………...…………...................................21

3 CHALLENGES OF GAME DEVELOPMENT ………..…………..…...…………….26

METHODS ...........................................................................................................26

AUTHOR INTRODUCTION AND BACKGROUND ........................................28

LEARNING HOW TO USE TWINE …………………………………………...32

INCEPTION OF DIVISION QUEST (A.K.A ORIGINAL DIVISION QUEST) ..34

TRANSITIONING TO A THESIS PROJECT ………………………………….39

STARTING FROM SCRATCH ……………………..…………………………41

SCRIPTWRITING ……………………………………………………………...49

PLAYABLE DEMO ……………………………………...……………………..52

ADDING CSS …………………………………………………………………...55

COMPLICATING FACTORS …………………………………...……………..64

ADDING IMAGES AND MOBILE-FRIENDLINESS ………………………...65

ADDING ILLUSTRATIONS AND FINAL TOUCHES ……………………….72

SUMMARY AND FINAL REFLECTION ……………………………………..75

4 DISCUSSION ……………………………………………...…...……………..............76

SUMMARY OF QUANTITATIVE RESULTS ………………………………...76

INTERPRETATION OF QUANTITATIVE RESULTS BASED ON

AUTOETHNOGRAPHY ………………….……………………...76

NEXT STEPS FOR DIVISION QUEST ………………………………………...77

LIMITATIONS ………………………………………………………………….78

5 CONCLUSION ………………………………………………...……………...............79

REFERENCES ....................................................................................................................82

APPENDIX ………………………………………………………….................................89

LIST OF TABLES

Page

Table 1: Five-question pre/post test used to assess mitosis learning …........................................89

Table 2: Demographic information of Georgia Southern University, Fall of 2022 ......................90

Table 3: Demographic characteristics for experimental group participants …….........................91

Table 4: Testing Assumptions for Comparison of Content Knowledge Between Instructor

Activities and Division Quest ……………………...…………………………………………….91

Table 5: Test statistics for Comparison of Content Knowledge Between Instructor Activities and

Division Quest ...…………………....................................................................................92

Table 6: Testing Assumptions for Comparison of Content Knowledge Between Between In-Class

and Out-of-class ………………………………………………………………………………….92

Table 7: Test statistics for Comparison of Content Knowledge Between In-Class and

Out-of-Class……………………………….......................................................................93

Table 8: Cross tabulation comparing game completion across learning contexts ........................93

Table 9: Testing Assumptions for Comparison of Content Knowledge Between Students Who Fully

Completed Division Quest and Students Who Never Played Division Quest …………………...94

Table 10: Test Statistics for Comparison of Content Knowledge Between Students Who Fully

Completed Division Quest and Students Who Never Played Division Quest...................25

Table 11: Summary of original challenges and decisions ……………........................................39

Table 12: Summary of brainstorming challenges and decisions …………...….………..……....48

Table 13: Summary of scriptwriting challenges and decisions ………...……………………….52

Table 14: Summary of first playable demo challenges and decisions …………………………..55

Table 15: Student Responses to “What feature(s), if any, did you find most fun or

enjoyable?” ……………………………………………………………………………...60

Table 16: Student Responses to “What feature(s), if any, did you find most helpful for your

learning?” ………………………………………………………………………………..61

Table 17: Student Responses to: “What parts of the game, if any, made the game less fun and

enjoyable and/or were not helpful for your learning?” ………………………………….62

Table 18: Student Responses to: “Would you suggest any changes that would make the game

more enjoyable for you or better help your learning?” ………………………………….62

Table 19: Summary of CSS addition challenges and decisions ………………………………....64

Table 20: Student responses to: “What feature(s), if any, did you find most fun or

enjoyable?” ……………………………………………………………………………...67

Table 21: Students responses to: “What feature(s), if any, did you find most helpful for your

learning?” ……………………………………………….………………………………68

Table 22: Student responses to: “What parts of the game, if any, made the game less fun and

enjoyable and/or were not helpful for your learning?” …………………………………69

Table 23: Student responses to: “Would you suggest any changes that would make the game

more enjoyable for you or better help your learning?” …………………………………70

Table 24: Student responses to: “Was the game ever confusing or hard to follow? If so, what

part(s)?” …………………………………………………………………………………70

Table 25: Summary of adding images and mobile-friendliness challenges and decisions ……...72

Table 26: Summary of adding illustrations and final touches challenges and decisions ………..75

LIST OF FIGURES

Page

Figure 1: Comparison of Content Knowledge Between Instructor Activities

and Division Quest ...………….............................................................................22

Figure 2: Comparison of Content Knowledge Between In-Class and Out-of-class ...........24

Figure 3: Screenshot of the Twine interface, showing its branching narrative structure ....33

Figure 4: Flowchart of conceptual design with handwritten notes from collaborative

brainstorming………….……................................................................................42

Figure 5: Sketch of UI concept ...........................................................................................46

Figure 6: Screenshot of a set of Twine passages ………………………………………….53

Figure 7: Screenshot of two Twine template listings on itch.io …………………………..56

Figure 8: Character portrait of Mat, the hematopoietic stem cell pedagogical agent …….73

Figure 9: Character portrait of a centrosome ……………………………………………..73

Figure 10: Two character portraits of the sister chromatids ………………………………74

11 
 
 
CHAPTER 1 
INTRODUCTION 
Difficulties and Misconceptions in Mitosis Learning 
Cell cycle and division are foundational concepts that must be mastered to understand 
higher level concepts, such as growth, development, reproduction, evolution, and genetics 
(McDonald & Gomes, 2013). Despite instruction during primary, secondary, and postsecondary 
education, evidence and the experience of many introductory biology instructors show that 
students’ understanding of the cellular process of mitosis is incomplete. Students may not grasp 
important events during interphase, such as the duplication of organelles (Ozcan et al., 2012) and 
often do not consider the fact that the genome is identical in all cells; they may think that a 
daughter cell has half of the genetic information of the parent cell, or that different cell types 
within the same multicellular organism have different genes relating to their function 
(Aldahmash & Alshaya, 2012). They may not fully grasp the relationship between genes, 
chromosomes, and cell division (Aldahmash & Alshaya, 2012; Ozcan et al., 2012) or that growth 
at the organismal level is synonymous with cell division at the cellular level (Aldahmash & 
Alshaya, 2012; Riemeier & Gropengießer, 2008). Commonly, they cannot track the movement of 
chromosomes during each phase of mitosis (Aldahmash & Alshaya, 2012; Dikmenli, 2010; 
Riemeier & Gropengießer, 2008). 
To remedy this lapse in understanding, it is important to understand why such a lapse 
exists. The cell cycle occurs at a microscopic level, which is abstract and far removed from 
students’ bodily experiences, obstructing their ability to integrate it into their existing mental 
models (Demirci, 2003; Dikmenli, 2010; Ozcan et al., 2012; Stepans, 2003). Additionally, 
students must navigate a suite of unfamiliar scientific terminology, many of which are 

12 
 
 
phonetically similar, such as chromosomes, chromatids, and chromatin (Lewis et al., 2000). 
Some of these terms were likely introduced during K-12 education; however, K-12 education 
emphasizes memorization of facts without linking concepts (Ozcan et al., 2012) at a variety of 
levels and ages, fragmenting concepts in a way that also inhibits construction of effective mental 
models (Chattopadhyay, 2005; Lewis & Wood-Robinson, 2000). In some cases, it may be 
introduced at too early an age, before students are mentally equipped to grasp abstract concepts 
(Chattopadhyay, 2005; Demirci, 2003; Stepans, 2003), and some have noted that many K-12 
science educators lack cellular and molecular expertise, which limits their ability to support 
mitosis learning (Klymkowsky, 2010).  
These misconceptions affect students’ ability to learn by disrupting their ability to 
correctly link concepts. To identify and correct these misconceptions, which is a vital step in the 
learning process, postsecondary instructors must select appropriate pedagogical techniques that 
do not reinforce misconceptions (Dikmenli, 2010; Ozcan et al., 2012; Rahma et al., 2022). 
Educators at all levels, including R1 institutions, tend to favor lecturing (Demirci, 2003; Stains et 
al., 2018; Stepans, 2003), which evidence suggests is insufficient to fill the gap in most students’ 
understanding (Freeman et al., 2014; H. G. Schmidt et al., 2015). Active learning, where students 
participate directly in learning activities, such as clicker questions, worksheets, call-and-response 
questions, and physical models or demonstrations, has been shown to increase student content 
knowledge and academic retention, but can be difficult to maintain individualized feedback with 
large class sizes (Cleveland et al., 2017). 
Games for Learning 
A novel form of active learning and growing research interest is games for learning. 
Games can take on many forms, from tabletop games to full-scale three-dimensional 

videogames. Due to this diversity, games (and by extension, games for learning) are difficult to

define without excluding something that is generally considered a game. For example, an avid

videogame player may define games as an interactive program played on a console or computer

where the player must take in information and execute actions to attain a specified goal, but this

definition would exclude tabletop games such as board and card games. Salen & Zimmerman’s

(2003) definition is widely cited and defines a game as a system having artificial conflict (such

as narrative conflict and/or competition between players), rules, and a quantifiable outcome.

Szilas & Acosta (2011) agree that games have an end goal (but do not mention either of the other

requirements) and highlight that player actions do not have consequences outside of the game.

Rather than giving a definition, Mayer & Estrella (2014) list some broad characteristics of

games:

● They have rules that define and structure gameplay

● They are responsive, meaning that players have some degree of autonomy in decision-

making and receive feedback on or responses to their actions

● They are challenging, requiring the player to exercise skills to achieve their or the

game’s goals

● They have cumulative progress, meaning that previous actions have continued effects

on the game experience

● They are engaging, motivating the player to interact and persist toward achieving their or

the game’s goals

For the purposes of this study, we define a game as a system that presents the player with a

challenge, gives clear rules that restrict or direct gameplay while still allowing the player to make

decisions that affect gameplay in observable ways, leading them to some form of resolution.

14 
 
 
Games for learning, then, are systems that meet these criteria and are intended to encourage, 
enhance, or facilitate learning, and they fall into one of three categories: gamification, playful 
learning, and game-based learning. 
  Gamification is the incorporation of game characteristics into a pre-existing learning 
activity without fundamentally altering the learning activity, where students are incentivized to 
complete the learning activity through achievements, points, rewards, or competition (Plass, 
Homer, et al., 2020). Examples include reward systems that are independent of any game-like 
activities (such as earning points on a leaderboard for completing learning tasks that are 
otherwise uninteresting) and tools such as Quizlet which allow students to make flashcard sets 
and earn scores based on how quickly they can correctly match terms or fill in blanks. 
  Playful learning is the alteration of a pre-existing learning activity using some, but not all, 
characteristics of a game to make the activity more interesting and facilitate learning. Playful 
learning assumes that a full game is not always necessary, but certain game features are useful to 
complement and enhance the learning experience (Plass, Homer, et al., 2020). This includes tools 
such as Kahoot and other game-show styled activities (Jeopardy, Wheel of Fortune, etc.) where 
the original learning activity has been incorporated into a new structure or environment with 
game elements, but does not require a complete redesign of the original learning activity (Jones 
et al., 2019). 
In contrast to the previous types of games for learning, game-based learning (GBL) 
requires the creation of an entirely new system or structure with which the learning activity has 
been intertwined. Rather than the learning activity being the primary focus of the otherwise 
unrelated system, it becomes a feature of it, allowing students to engage with the material 
indirectly through their exploration of the system (Plass, Homer, et al., 2020). GBL has been 

15 
 
 
explored in many forms in multiple disciplines. Puzzle games have been used in computer 
science, engineering,  and physics, such as Circuit Game, where players learn about electrical 
circuits and Ohm’s Law (Johnson & Mayer, 2010; Mayer & Johnson, 2010). Simulations, both 
of patient interactions and medical procedures, allow students of medical, nursing, and pharmacy 
school to practice and demonstrate skills in a safe way before beginning clinical work (Litten & 
Stewart, 2023; Morningstar-Kywi & Kim, 2021). For a comprehensive review of GBL research, 
we recommend the Handbook of Game-Based Learning (Plass, Mayer, et al., 2020). 
Commercially available GBL tools tend to be expensive, with costs falling either on the 
instructor or the students (Watson & Yang, 2016). Others may be more affordable, but are less 
applicable to learning objectives due to being designed and marketed more broadly for 
entertainment purposes than specifically targeting an academic audience. While these 
commercial entertainment products may be high in visual and narrative quality and offer an 
engaging experience, they seldom can be modified to better align with individual course learning 
objectives (Lester et al., 2023). Even those that do make assets available for customization 
require high level programming skills to do so, which many instructors lack and do not have the 
time or resources to attain (Lester et al., 2023; Watson & Yang, 2016). Previous studies have 
shown that GBL tools that have significant effects on learning tend to use independently created 
games designed to align with learning objectives (Anguera et al., 2013; Homer et al., 2018; 
Litten & Stewart, 2023; Morningstar-Kywi & Kim, 2021; Parong et al., 2017). 
However, developing a GBL tool oneself need not necessarily require extensive 
programming skills or expensive software and training. Free and open-source software (FOSS) is 
available to remove the financial cost of entry, and some FOSS platforms also utilize intuitive 
and beginner-friendly user interface (UI) and markdown language to facilitate use by novice 

16 
 
 
programmers. One such FOSS is Twine, an HTML-based programming platform primarily used 
to create interactive fiction (IF), a form of branching narrative where reader (or player) choices 
influence the story and lead to different outcomes, based on and functionally similar to the 
Choose-Your-Own-Adventure books of the 1970s. IF is primarily text-based and accepts user 
input to affect the narrative, either through parsing prompts, where users make choices by typing 
in their desired input, or through hyperlinked text, where users choose from a finite number of 
given options, which is the more common choice among Twine developers (Terry & 
Dusenberry, 2018).  
Twine’s software bundles the three pillars of web development: HTML (Hypertext 
Markdown Language), CSS (Cascading Style Sheets), and JavaScript; and provides a simple 
interface with which to interact with them (this will be discussed in detail in Chapter 3). Twine’s 
homepage, where the free desktop app download is available, declares, “You don’t need to write 
any code to create a simple story with Twine, but you can extend your stories with variables, 
conditional logic, images, CSS, and JavaScript when you’re ready” (Twine / An Open-Source 
Tool for Telling Interactive, Nonlinear Stories, n.d.). This can-do attitude extends beyond the app 
developer into the Twine community, where Twine users eagerly share their knowledge and 
resources with one another on various social media and digital game market platforms. Twine’s 
ease of use, versatility, and mutually supportive FOSS community allow anyone, regardless of 
programming experience, to create simple IF stories on any topic that can be modified and 
distributed at no financial cost. 
Game-Based Learning Theoretical Frameworks 
Because GBL tools exist in many forms and styles, there are likewise a variety of 
theoretical frameworks that have been applied to examine GBL efficacy in different contexts. 

17 
 
 
Twine GBL primarily leverages its narrative as a learning tool; therefore, viewed through the 
lens of constructivism, which holds that knowledge is constructed through experiences that must 
be integrated into existing mental models based on prior knowledge and experiences, GBL can 
contextualize learning within life-like scenarios where learners can process, organize, and apply 
their knowledge in a space that is safe to make mistakes (Plass et al., 2009; Shaffer et al., 2005). 
By providing an interactive environment with the use of targeted feedback and scaffolding, GBL 
allows learners to accomplish difficult learning tasks without becoming frustrated, placing them 
directly in the zone of proximal development, where the learning activity is challenging enough 
to force the learner to build on and exercise use of their mental models but not so challenging 
that they become overwhelmed and stop engaging with it (Plass et al., 2009; Vygotsky, 1962). 
Additionally, overcoming these learning obstacles without excessive struggle helps learners 
reduce science anxiety and academic self-efficacy (the belief in their ability to succeed 
academically and meet learning goals), which are both positively associated with student success 
(Bryant et al., 2013; England et al., 2019; Mallow & Greenburg, 1983)). . 
However, narrative as a learning tool is a double-edged sword; while some immersion in 
the narrative has been shown to promote learner motivation and positive affect, and therefore 
facilitate learning, evidence suggests that high levels of immersion detract from learning due to 
increased cognitive load, the demand on the brain to process stimuli and organize or discard 
information from it. A hybridized theory of GBL, Mayer’s (2020) cognitive theory of GBL, 
adapted from Mayer’s (2009) cognitive theory of multimedia learning and combined with 
Sweller et al.’s (2011) cognitive load theory, states that the visual and auditory stimuli provided 
by the game are processed in sensory memory, where information that the learner deems 
irrelevant or less important are filtered out. Therefore, it is important to prioritize the stimuli a 

GBL tool provides; germane load (cognitive load directly related to learning objectives) must be

emphasized, while extraneous load (cognitive load not directly related to learning objectives)

must be minimized (Sweller et al., 1998).

Research Questions

The present study seeks to investigate the use of an independently-developed Twine

game for the purposes of improving student learning outcomes in relation to the concepts of

mitosis in an introductory level biology course and overcoming some major barriers to

implementing GBL. We therefore have two sets of research questions: one related to

understanding the efficacy of a GBL approach for mitosis learning and the other relating to

understanding the challenges of developing this type of GBL tool and modeling strategies to

overcome them.

Learning Question 1: Are there differences in student mitosis knowledge between

students who used the GBL tool and students who completed instructor activities?

Learning Question 2: Are there differences in student mitosis knowledge between

students who completed the GBL tool in-class and students who completed it out-of-

class?

Learning Question 3: Are there differences in student mitosis knowledge between

students who completed the GBL tool and students who did not use the GBL tool?

Development Question 1: What are the challenges of making a GBL tool for mitosis

learning in Twine?

Development Question 2: What strategies for overcoming these challenges were

effective, and why?

CHAPTER 2

EFFICACY OF GBL APPROACH

Methods

Because it is best practice for GBL tools to align with instructor style and specific course

learning objectives, we intended to recruit study participants from introductory biology lecture

sections taught by Dr. Sue Ellen DeChenne-Peters, as the game was being designed for her

introductory course. Therefore, to assess student learning, we designed a five question pre/post

test (shown in Table 1 in the Appendix on pg. 90) based upon her mitosis learning objectives.

We selected learning objectives that we wanted the game to target, keeping in mind the

limitations of Twine and my own programming ability, then located questions in her question

bank that aligned with each learning objective and identified their Bloom’s taxonomic level

(Benjamin S. Bloom, 1984). We selected questions that targeted the most learning objectives

overall and hit an even spread of Bloom’s levels (2 remembering level questions, 2

understanding level questions, and 1 application level question). Data was to be collected

immediately before and after learning intervention and comparisons made between students

completing the instructor’s standard activity and those using the GBL tool.

Unfortunately, early in the project, Dr. DeChenne-Peter’s teaching schedule shifted and

no longer included in-person introductory biology lecture sections, so we sought to collect data

with other instructors. Due to a recent consolidation of Georgia Southern University (GSU) with

Armstrong State University, some course curriculum differs between the Statesboro and

Armstrong campuses. On the Armstrong and Liberty campuses, cell division is covered in

lecture sections and student learning outcomes (SLOs) include identifying the reason cells

divide, defining terminology, relating DNA replication to chromosomal structure, identification

of key processes phases of the cell cycle, identification stages and sequence of events in mitosis,

identification of the stage when chromosomes condense, and identification of the function of

centrosomes. On the Statesboro campus, cell division is covered in lab sections and SLOs

include visual identification of mitotic phases under a microscope and identifying the purpose of

division. Because the game was designed based upon the learning objectives of an Armstrong

instructor (Dr. DeChenne-Peters), our goal was to recruit only from sections offered on the

Armstrong and Liberty campuses. However, circumstances limited our available data collection

to a single instructor on the Armstrong campus teaching two in-person lecture sections, one of

which was an honor’s section. To avoid an exceedingly small sample size and potentially

skewing our data by including the honor’s section, we extended our recruitment efforts to the

Statesboro campus labs.

Study participants were recruited from students taking an introduction to cellular and

molecular biology course with an associated lab at Georgia Southern University. Due to limited

lab time, participants in Statesboro sections completed the game as an out-of-class pre-lab

assignment, with the pre-test taken in-person via Folio at the end of the lab meeting the week

before mitosis was to be covered (when the mitosis pre-lab assignment was assigned) and post-

tests taken at the beginning of lab meeting (when mitosis was to be covered). Armstrong

participants completed the game as an in-class active learning activity, with pre- and post-tests

taken outside of class via Folio within 24 hours before and after intervention. The recruitment,

intervention, and assessment methodology was approved by and in compliance with the

Institutional Review Board (IRB) of Georgia Southern University (Protocol #H23057).

Georgia Southern University is an R1 institution of around 25,000 students as of 2022; of

those, 1,693 students (6.6% of the university’s total enrollment) were enrolled in the College of

Science and Mathematics, with 1,070 of those pursuing a B.S. in Biology. General demographic

information for Georgia Southern University is shown in Table 2 in the Appendix (pg. 91,

Georgia Southern University Fact Book 2022-2023, n.d.).

Demographic data, including race and gender (not sex), were collected alongside pre-test

scores and informed consent within the Division Quest condition. They were not collected within

the control condition due to miscommunication. This was done with the intention of performing

additional analyses based on race and gender to observe potential differences in GBL efficacy for

underrepresented and marginalized groups, primarily women and students of color. However,

due to lack of demographic data in the control condition and small sample size, these analyses

were not performed. Descriptive statistics of the sample are shown in Table 3 on pg. 91 of the

appendix.

Data were analyzed utilizing MANCOVA, ANCOVA, and Chi-square statistical tests as

appropriate using IBM SPSS Statistics Version: 29. Pre and posttest data were matched on

student name and ID. For all analyses, significance was determined using ɑ = 0.05.

Results

To determine if there are differences in student mitosis knowledge between students who

used the GBL tool and students who completed instructor activities, we conducted a one-way

MANCOVA comparing posttest scores between learning conditions (GBL and instructor

activities) while controlling for learning context (in- and out-of-class) and pretest scores. For the

total score, we ran a one-way ANCOVA using the same variables. MANCOVA results show that

there is a significant difference in post-test scores between students who completed instructor

activity and those who played Division Quest on Question 1: On Duplicated Chromosome

Appearance, with students who played Division Quest scoring higher than those who participated

in instructor activities (p = 0.024, Figure 1), and Question 4: On Recognizing Phase, with

students who participated in instructor activities scoring higher than those who played Division

Quest (p = 0.003, Figure 1). For all other test questions and the total score, there was no

significant difference in posttest score between students who played Division Quest and those

who completed an instructor activity. Results are shown in Figure 1, and test assumptions and

test statistics for all questions and total score are shown in Tables 4 and 5 in the Appendix (pg.

91-92).

Figure 1: Comparison of Content Knowledge Between Instructor Activities and Division Quest

Note. This figure shows the estimated marginal means of post-test scores between students who

completed instructor activity and those who played Division Quest. Error bars represent the

standard error of the mean. * p < 0.05

To determine if there are differences in student mitosis knowledge between students who

completed the GBL tool in-class and those who completed it out-of-class, we conducted a one-

way MANCOVA comparing posttest scores between learning contexts (in- and out-of-class)

while controlling for pretest scores. For the total score, we ran a one-way ANCOVA with the

same variables. MANCOVA results show that there is a significant difference in posttest scores

between students who completed Division Quest in-class and those who completed Division

Quest out-of-class on Question 3: On Splitting of Duplicated Chromosomes, with students who

completed Division Quest in-class scoring higher than students who completed it out-of-class (p

= 0.001, Figure 2), and total score (p = 0.047, Figure 2). For all other test questions, there was no

significant difference in posttest score between students who completed Division Quest in-class

and those who completed it out-of-class. Results are shown in Figure 2 and test assumptions and

test statistics for all questions and total score are shown in Tables 6 and 7 in the Appendix (pg.

92).

Figure 2. Comparison of Content Knowledge Between In-Class and Out-of-class

Note: This figure shows the estimated marginal means of post-test scores between students who

completed Division Quest in-class and those who completed Division Quest out-of-class. Error

bars represent the standard error of the mean. *p < 0.05

To determine if there are differences in student mitosis knowledge between students who

completed the GBL tool and students who did not use the GBL tool, we first needed to determine

the relationship between learning context and completion level. To do this, we conducted a Chi-

square test of independence (shown in Table 8 in the Appendix on pg. 93), and the relationship

was found to be significant (p <.001). Based on these results, only the Statesboro responses from

the Division Quest semester were selected (being the larger of the two subsamples with a greater

balance between students who fully completed the game and students who never played the

game) to conduct an ANCOVA using completion as a variable. All respondents who reported

partial completion were removed due to the wide range of possible game exposure. We then used

the remaining data to conduct a one-way MANCOVA comparing posttest scores between

completion levels (fully completed and never played) while controlling for pretest scores. For the

total score, we ran a one-way ANCOVA with the same variables. Results showed that there were

no significant differences in posttest score after controlling for pretest score between students

who completed Division Quest and those who never played Division Quest (shown in Table 10).

To determine if these results were due to having too small of a sample size to detect a small

effect size, we performed a post hoc power analysis (also shown in Table 10) which determined

that our observed power was low (<0.8). Therefore, it is possible that there was a small

difference in effect size that we were unable to detect due to insufficient sample size.

Table 10: Test Statistics for Comparison of Content Knowledge Between Students Who

Fully Completed Division Quest and Students Who Never Played Division Quest

EMM

SEM

EMM

SEM

b/t

w/i

Observed

Power

0.290

0.090

0.359

0.0698

0.349

0.557

0.090

0.493

0.098

0.571

0.075

0.379

0.540

0.093

0.501

0.097

0.410

0.074

0.529

0.470

0.111

0.499

0.099

0.367

0.076

1.052

0.309

0.173

0.300

0.097

0.442

0.074

1.273

0.263

0.199

Total

0.422

0.055

0.427

0.042

0.004

0.949

0.060

Note: EMM = estimated marginal mean and SEM = standard error of the mean, A = Fully

completed Division Quest, B = Never played Division Quest.

26 
 
 
CHAPTER 3 
CHALLENGES OF GAME DEVELOPMENT 
Methods 
Autoethnography is a qualitative research method in the genre of ethnographic research (Bochner 
& Ellis, 2022). It utilizes reflection on personal experiences as a means to understand cultural or social 
phenomena via description and analysis, contextualized by existing literature. Researchers in a variety of 
fields, including social sciences (Bochner & Ellis, 2022), education research (Keleş, 2022), and applied 
ecology (Murphy et al., 2022) use this method not only to analyze, but also to convey details of how a 
project was undertaken that are often minimized in favor of objectively describing and reporting 
quantitative approaches in writing intended for scientific publications (Keleş, 2022; Murphy et al., 2022). 
For example, a conservation scientist can follow the methods of a previously published paper to conduct a 
population survey from the methods of published studies; however, that same scientist may not have the 
knowledge of how to interface with one or more levels of government (local, county, state, national, etc.) 
or private landowners in order to gain access to the areas of interest. Using autoethnography allows 
scientific authors to share these important details that can give their readers a more complex, complete 
picture of their research. 
Researchers may also use autoethnography to present a more complete picture of their research or 
research subjects so that quantitative findings are contextualized by qualitative specifics that can help 
form deeper connections that potentially lead to deeper understanding (as in social sciences, Bochner & 
Ellis, 2022) Because autoethnographic research often uses a narrative writing style, it can make the topic 
more palatable, and even enjoyable, for readers in other fields of research or the general public who may 
not be able to or don’t want to put in the extra effort to parse through unfamiliar topics and terminology. 
When used in this way, autoethnography can be experienced more as a story, with the researchers as 
characters and the project and associated difficulties as conflict (Keleş, 2022) and can be leveraged to 
reach a broader audience and have impact beyond research settings. 

As described, autoethnography is a departure from methods that emphasize quantitative data,

statistical analysis, and removal from subjectivity. However, when used to draw connections between

decisions made and existing literature evidence, and in conjunction with quantitative methods, it allows

for a more holistic understanding of the phenomenon of study while facilitating interdisciplinary

collaboration and future endeavors (Murphy et al., 2022). In addition to answering our research questions,

an overarching goal of this segment of the project is to encourage and assist other higher education

instructors to develop GBL tools for their own courses that suit their unique teaching style and learning

objectives. For this purpose, we require an analytical approach and writing style that clearly presents the

challenges of doing this work while continuing to manage the other duties and responsibilities of being an

academic and, more importantly, ways in which those challenges can be overcome, which

autoethnography is uniquely suited to accomplish.

To answer our research questions relating to game development, we conducted an

autoethnography on the game development process from conception to final product, including

several iterative versions of the game with accompanying postmortem analyses. Postmortem

analysis, while not unique to game development, is often utilized in commercial game

development and release and involves description of and reflection on the development process,

decisions or events that influenced game development, goals throughout the project, whether or

not those goals were achieved, and reflection on successes and/or failures (Wirtz, 2023). Notes

were kept via a Word document changelog that included decisions made (either based upon the

literature or in response to challenges faced), documentation of thought processes during

decision making, and documentation of mental and emotional state through comments attached

to notes and screenshots of conversations with friends in Discord, a social media platform

primarily used by gamers and game developers. At the end of the project, the autoethnography

narrative was written by identifying and discussing the major challenges and decisions within the

notes and giving a final reflection with reference to findings from statistical analyses detailed in

Chapter 2.

In addition to changelog notes and Discord messages, student feedback was collected

based on two preliminary versions of the game. Because this feedback was intended only to

inform the development of the final version of Division Quest used to collect data for statistical

analysis detailed in Chapter 2 and not for direct analysis, feedback was categorized based upon

game features. Example quotes were chosen where either the most influence on future decision

making and/or most unanticipated based on my own perception of the game at the time.

Author Introduction and Background

Because everyone’s stories and experiences are unique, what works for me may not be

applicable to other practitioners. Additionally, factors that I may or may not be fully aware of,

such as the biases of myself and others involved in my story, play a significant role in the content

of my experiences and how I perceive them. For these reasons, Murphy et al. (2022) asserts that

to gain meaning from the experiences of another, the reader must understand the author’s

background and the context surrounding their experiences, especially potential biases. Therefore,

I’ll start by giving you a brief overview of my background and demographics and share with you

my personal history with gaming, GBL, and how I found myself in the position of developing

and researching GBL tools.

I’ve lived my entire life in the Savannah, GA area, with my childhood spent in a small

town where everyone knew my parents, grandparents, and great-grandparents. I had a varied and

rigorous primary and secondary education (including private elementary school, homeschooled

in middle school, and public high school with mostly AP/Honors courses). I was fortunate to

meet my now-husband early in life, and he was able to support me through my bachelor’s degree

so I could afford to take my time as an undergraduate and still be student-loan free. I’m white,

female-presenting, nonbinary, and, at the time of writing this, twenty-seven years old and the

mother of a curious, outgoing 9 month old.

I consider myself an avid gamer and have for as long as I can remember. Whenever I had

pocket money as a kid, you could find me pacing the game display case in Walmart trying to

decide which game to buy. Playing videogames, both alone and with friends, made up many of

my fondest childhood memories and stoked my curiosity. My mother noticed my sponge-like

knowledge retention when it came to games I enjoyed and realized that she could leverage that

phenomenon to give me an educational head start by giving me Hooked on Phonics: Learn to

Read on her old Dell Dimension. The colors, characters, sounds, and interactivity captivated me,

enabling me to develop my lifelong love of reading. This trend continued throughout elementary

school, with Type to Learn, Funbrain, Land Before Time Activity Center, Pokemon Masters

Arena, and many others making appearances on both my home computer and in the school

computer lab.

However, from middle school on, education and gaming diverged in opposite directions,

with school being centered around textbooks and lectures and games being solely for

entertainment. That being said, playing games still allowed me to exercise many skills that I use

professionally on a daily basis: self-reflection and problem solving when I fall short of my goals

(both assigned by the game and self-motivated); pattern recognition and attention to detail (ex:

spotting the “big glowing weak spot” on boss enemies, which are enemies that are more

challenging and often require execution of specific actions to defeat, such as targeting an

illuminated part of the boss’s body); and interpersonal skills, such as leadership, clear

communication, and conflict resolution (ex: leading a group of online gamers in a coordinated

attack, known as a raid, that can only be challenged as a group in which each member is prepared

with knowledge of the enemy and its attacks, strategies of how to evade attacks while leveraging

counterattacks, and making split-second adjustments while under high emotional tension).

I’m also a first-generation college student and had what I consider to be an atypical

college experience. I came out of high school with a large amount of credit from my AP exams,

putting me significantly ahead in my core requirements. I progressed through the general

chemistry requirements quickly and made it to organic chemistry much sooner than I otherwise

would have. That was the first time in my life that I almost failed a course, and the first time I

was regarded as A Woman Who Couldn’t Make It In STEM. This external and internal pressure

pushed me to the edge, and I ended up changing my major.

I was undeclared for a time while I re-evaluated and eventually declared nursing as my

major. It was in nursing school that I discovered my passion for teaching. I loved speaking with

patients and meeting them at their level to help them understand the biology of their condition

and the pharmacology of their treatment plan. Ultimately, I left nursing school, only one

semester short of a degree, due to an acute medical condition (ironic), but the doctors, nurses,

PCTs, and nursing instructors that I learned from showed me that my view of science was too

narrow; scientists are varied and diverse people who apply their knowledge and skills in many

different ways. Most importantly, I learned that being able to explain complex ideas, especially

to those with little or no science background, was a valuable skill and one at which I excelled.

Those experiences set the foundation for the work I would do as an undergraduate and, by

extension, the current project.

Then COVID happened.

One benefit of my first semester as a returning biology major being online: all of my

morning routine and commuting time was suddenly free for other activities. Most days, I spent

that time gaming. My favorite game at the time was Beyond Blue, a singleplayer, story driven

narrative exploration game about marine science research. You play as Mirai Soto, a deep-sea

diver and scientist working as part of a team. Each researcher in the game is studying something

different: Mirai is documenting sperm whale behavior, primarily child rearing; Irina investigates

coral bleaching and antibiotic properties; and Andre collects recordings of whale songs and

explains that he hopes to identify regional dialects. Ultimately, he plans to conduct an

experiment where he plays the recorded whale songs from one region for whales in a different

region and determines if the whales respond differently to foreign dialects than their own.

When I listened to that, I lost my mind. Earlier that same day, I watched a video lecture

in my evolution and ecology class on the types of isolation that can lead to speciation, including

behavioral isolation: when the mating practices and preferences of individuals within a species

lead them not to choose each other as mates and, therefore, prevent them from producing

offspring. That’s exactly what Andre was describing! I was ecstatic to see a concept from one of

my classes brought to life. I could see it in practice! With my own eyes! That was the first time I

had considered the overlap between my formal science education and my informal gaming

hobby. To this day, I recommend Beyond Blue to any colleagues that teach evolution or marine

science courses or are interested in my GBL research.

Fortunately, I made that recommendation to one of my instructors the following semester

when classes resumed meeting in-person, and rather than just jot down the title and assure me

she would look into it, she mentioned that this idea might interest another faculty member at

Georgia Southern (now my mentor and co-author, Sue Ellen DeChenne-Peters) and helped me

arrange a meeting. Unfortunately, Sue Ellen had immunocompromised family at home and was

teaching remotely. We had to meet over Zoom, and she wasn’t in a place to take on a new

research focus at the time, but she did encourage me to continue pursuing this interest, and that’s

exactly what I did.

Learning How to Use Twine

Not to be deterred, I continued collecting games whose content overlapped with my

science courses (for example, Niche – a genetics survival game, Alba: A Wildlife Adventure, and

Bee Simulator) while also dipping my toes in game development by taking a digital storytelling

course about making interactive fiction in Twine, taught by my now co-author, Rob Terry. I

consider that class to be the greatest influence that pushed me toward GBL research. In the

course of a single semester, I went from clueless about programming to making complete games

on my own. One day near the end of the semester, I was browsing my practice files to see what

features I liked and wanted to explore more in my next game that would serve as the course’s

final “exam”. I saw something that looked like Figure 3, which shows how Twine displays story

passages for editing. Each box is a passage, labeled with its title, and the arrows signify possible

paths that the player can take through them. This branching narrative structure is the hallmark of

interactive fiction, but for the first time I noticed that it looked similar to something I’d seen in a

microbiology lab: a dichotomous key.

Figure 3. Screenshot of the Twine interface, showing its branching narrative structure.

I took a break from brainstorming my game to reminisce on how much fun that lab was.

The bulk of it was an unknowns project, where we each took an unknown microorganism and

performed a series of tests on it to identify the species using a dichotomous key. My classmates

and I were so nervous when the project was introduced, because of it making up a significant

portion of the course grade and the instructor emphasizing that if we made a mistake and took an

incorrect path on the dichotomous key, we would land on an incorrect identification and there

may not be time left to correct it. However, by the end of the semester, we were much more

confident in our abilities and laughed with each other over how needlessly worried we were. If

only there were a way we could have seen how easy the tests on the dichotomous key are, I

thought.

Oh! What if I made it into a game?!

After coordinating with both instructors, that’s exactly what I did. The microbiology lab

instructor shared lab materials with me, including the dichotomous key, lab protocols, and photos

of lab equipment and all necessary cell cultures and test results, and Rob gave me guidance on

how to maintain the integrity of the lab procedures within an engaging, interactive narrative. By

the end of the month, I’d converted the lab project into a text-based narrative simulation,

complete with random selection of 1 of all 28 possible microbes in the lab and a pedagogical

agent in the form of a virtual reality AI teacher.

During the same time, I secured a part time job as a learning assistant (LA) for an

introductory biology course. I would work closely with the faculty member I had been assigned,

supporting the students in and out of the classroom; I had been assigned to none other than Sue

Ellen, the professor I had met with a year prior about my interest in the utility of games for

learning! She actually turned me away when I went to her office to introduce myself as her new

LA. “I’m expecting another student soon, can you come back later?” she said. Later, after we’d

become close colleagues and friends, she asked me if I remembered that happening. I told her

yes. “I didn’t know you were my new LA!” she laughed. “That’s who I was expecting, and I

didn’t know it was you!”

Inception of Division Quest (a.k.a Original Division Quest)

I supported Sue Ellen’s Introduction to Principles of Biology I course for a year. Because

she emphasizes active learning, a large part of my job was to walk around the room during class

observing the students at work and give guidance when needed. Usually, the students do well

working in groups and only need the occasional nudge. However, one day, while covering the

phases of mitosis, I noticed that I was spending much more time offering clarification and basic

instructions than usual. The activity was web-based (while most of Sue Ellen’s activities are on

paper), and the webpage design was… less than ideal. Cell images were miniscule, grainy, and

low contrast, which resulted in the students struggling just as much, if not more, with the tool

itself than with the concepts being learned.

“It’s not the best, but I think it’s important for them to see these real cell images,” Sue

Ellen explained when I voiced my frustration.

“They can’t really see them anyway,” I laughed, to which Sue Ellen shrugged.

“Until I find something better, this is what I have.”

“...I could probably make something better.”

Sue Ellen’s eyes went wide. “Could you???”

We spent our next meeting going over the concepts she wanted emphasized, and I took

notes on how I could integrate those into a game. The original vision was for the students to

choose what kind of cell they would be (either plant, animal, or fungus), and they would proceed

through the cell cycle as somewhat of a manager of the cell. In terms of game genre, it would

most resemble a resource management game, with the player given various tools to manipulate

the environment and perhaps collect some form of in-game currency or achievements. Notably,

this version of the game would have the player taking on a much more detached role in the

narrative with very little emotional investment. This would become a problem later as I began to

take the game more seriously and explore GBL literature (more on that in “Starting from

Scratch: Division Quest Version 1”).

At this stage, the top priority was getting better cell images. I’d learned from previous

games that it’s usually easier to take or draw images myself rather than find appropriate images

somewhere else that I’m legally allowed to use and make sure to properly cite them. Sue Ellen

suggested I reach out to our lab coordinator to get access to a microscope with a camera

attachment and prepared slides of cells undergoing mitosis. Our coordinator has always been

highly invested in our students, particularly students doing undergraduate research (which, at the

time, was functionally what I was doing). When I told her what I needed and why, she was

ecstatic.

“I have the perfect cells for you!” she said with a grin on our way to the lab where the

slides were stored. “Fish blastula. Way prettier than the onion root tips we usually show

students.”

When I saw them, I had to agree. The cells were stained bright pink with the

chromosomes a deep shade of purple, and the cytoskeletal elements were visible!

“You can see the mitotic spindles!! Why isn’t this what we show to students??? These are

amazing!”

“I knew you’d love them.”

There were a few reasons I chose these cells beyond them being recommended to me. For

one, they were animal cells, which was one of the three cell types I’d planned to include in the

game, and Sue Ellen decided in our meeting that of the three cell types, she wanted students to

see animal cells the most. At a later time, once the game was functionally complete, we could go

back and add plant and fungal cells to be displayed depending on what type of cell the students

picked, but for the first iteration, all cells looking like animal cells was fine with us. Secondly,

the mitotic spindles were visible, which I thought could mitigate the way that students often

overlook this vital component of mitosis.

I started working on coding the game in my spare time, but wasn’t really serious about it

yet. I didn’t even make an outline or paper prototype (where you write out, often on paper, how

you expect the game to look and/or function. I use flowcharts, usually); I just wrote directly in

Twine whatever ideas came to me in the moment. The first major hurdle was the part of making

Twine games that I always struggle with the most: making the visual elements of the game, the

Cascading Style Sheets (CSS). CSS is the part of a web program that dictates the visual display

of the program. Just like all programming languages have their own unique syntax (the format in

which you must write the code for the program to be able to recognize and execute it), CSS also

has its own syntax. Twine still uses its wonderful color-coding system to help you navigate the

CSS, but the syntax is much less intuitive than that within the story passages. Thankfully, the

Twine community is a FOSS community, and Twine users are dedicated to the collaborative

nature of storytelling and programming and believe everyone in the community should be

sharing our code to enable us to grow together. Because of this, finding an open source template

similar to what I want to use as a starting point is my preferred method for visually upgrading my

Twine games. I find it much easier and less intimidating to modify an existing template than to

try to build something from scratch.

Unfortunately, as I progressed with making the game, I realized to my horror that the

template was for an old version of Harlowe. Harlowe is one of several Story Formats available

for Twine. Story Formats are where all the magic of Twine happens; they take your story from

the Twine application and convert it into a webpage where the user can interact with (a.k.a. play)

it. Of the Story Formats for Twine, Harlowe is generally considered the most user-friendly and

easiest to pick up with little or no programming experience. However, because of the low

programming skill needed, Harlowe isn’t capable of some of the more complex functionality that

HTML programs are capable of. SugarCube is another popular Story Format for Twine, which

unlocks a little more of that functionality with the caveat that more programming knowledge is

required. Others, like Snowman, remove the training wheels entirely and require the programmer

to manually build much of what is built-in to Harlowe and SugarCube, but lose the restrictions

associated with them.

The template being in an older version wouldn’t have been a problem, as the game would

still be perfectly functional regardless of what Story Format or version it was made in. The

problem was that the newest version of Harlowe had innate functions that I wanted to use that

didn’t exist in the older version the template existed in. Part of what makes Harlowe more

accessible than other Story Formats are its use of macros, which are a prepacked code function

meant to make it easier to utilize and display specific elements. Rather than build the code by

hand, the developer can simply use the macro shorthand to achieve the desired effect. In this

case, I wanted to use the (button:) macro, which turns the normal hyperlinked text into a

clickable button. This was partially a cosmetic change, because I love the way the buttons look,

and partially a potential cognitive load management tool. The larger, bolder button draws the eye

and guides the player’s attention to the decision to be made and, in theory, reduces extraneous

load.

Partially out of frustration over the template and partially because of the workload of my

final semester as an undergraduate, I tabled the game here for a little while. Because this was still

a little side project, I hadn’t delved into the literature very deeply and had no idea about

evidence-based practices in GBL development. With that in mind, what I made up to this point

was more a gamification of learning activities from Sue Ellen’s class than true GBL. Embedded

assessments, which were unaltered content questions, interrupted what little narrative there was,

and the characters were more set pieces than real characters that the player could become

emotionally invested in.

Table 11: Summary of original challenges and decisions

Challenges

Decisions

Existing learning tool fulfills

the learning objective needs but

is difficult to use and frustrates

the students (and us, a little)

Make a better one!

● take higher quality cell images and display them at a

larger size

● incorporate other learning objectives beyond visually

identifying phase of the cell and understanding how

much time the cell spends in each phase

Outdated CSS template

● missing a feature of the

newest version of

Twine that I wanted to

use, but difficult to port

to the new version

First attempt: reach out to the Twine community for help

updating the template

● successful, but template becomes obsolete as my

vision for Division Quest changes throughout my

literature review

Second attempt: start over on a blank slate

Transitioning to a Thesis Project

As I was nearing graduation, I started making plans for what was next. My experience as

an LA had reinforced my love of teaching and given me comfort and confidence in teaching at

the college level. Sue Ellen had already become a mentor for me, and because I knew her course

content and teaching preferences well, we spent most of our class prep time that last semester

discussing evidence-based teaching and the reasoning behind her decisions in course structure.

When I expressed interest in making science education my career, she heavily encouraged me

and gave advice on what kind of program to look for and how to reach out to university faculty

that I was interested in working with.

I looked at several programs along the southeast coast, because marine science has

always been my main area of interest within biology, and with my husband not wanting to leave

his job and both of our families in the immediate area, I didn’t want to go far. Each week at our

prep meeting, Sue Ellen would ask me how my search was going. I’d keep her updated on who

I’d reached out to and who had responded, but as fall turned into spring with nothing panning out

yet, I started looking for work to keep me going after graduation while I continued my search.

One day, after discussing that, Sue Ellen abruptly changed topics to my game.

“Are you still working on that mitosis game in your spare time?”

“Yeah, but I’m having some trouble on the coding side right now. My template is in an

older version that doesn’t have a new feature I want… long story short, yes.”

“Do you like doing that? It sounds like it’s rough.”

“Oh yes, I love making games, and I’m really interested in seeing how this specific kind

of game could make game-based learning more accessible.”

“That sounds like a master’s thesis to me,” she said.

“...A what?”

“A master’s thesis. Here. With me.”

“...Are you sure??”

“I mean, if you want to. You don’t have to.”

“NO, YES! YES I WANT TO!”

The grad school application deadline was very close, but I got my application package

prepared in time, and after I was accepted, Sue Ellen and I began discussing what my project

would look like so I could spend the summer drafting my thesis proposal and IRB application

and we could gather control data as soon as possible.

Starting from Scratch

After writing my thesis proposal, getting IRB approval, and selecting a more specific set

of learning objectives for the game to target, we created the pre/post-test described in Chapter 2.

With learning objectives selected and a working knowledge of evidence-based practices in GBL

development, I resumed game development. Well, ‘started over' would be more accurate.

Original Division Quest had almost no narrative and was really a loose tying together of various

pieces of active learning worksheets and in-class quizzes Sue Ellen had been using for years. It

wasn’t a GBL tool, and it wasn’t really fun unless you’re a massive nerd like me (which 99% of

students are not, and that number is really lowballing it, because I am a singularity of nerdiness).

In the revived version of Division Quest, I wanted the player character to have a more

active role in the narrative and, by proxy, the concepts of mitosis embedded in the narrative. I

started with a flowchart (shown in Figure 4) detailing the “correct” path for students to take and

attached notes where I had ideas for character interactions, game mechanics, and GBL features

from the literature. Within the flowchart, the game had already taken on its basic shape: the

player would start the game, get to choose what kind of cell they are, be introduced to the

characters and setting, and reach the inciting incident that kickstarted the narrative. That’s where

my logic stepped on the toes of my creativity. Stumped, I took my flowchart and a list of

questions to a lab meeting to share and discuss with Sue Ellen and our undergraduates.

Figure 4. Flowchart of conceptual design with handwritten notes from collaborative

brainstorming

“Why would a cell not know how to divide?” I remember asking.

“Because you need it not to,” Sue Ellen replied matter of factly.

“Well, ok, yeah, fair.”

“Amnesia,” one of our undergrads said. “It’s always amnesia.”

“Oh, we love a good cliche,” I laughed as I scribbled in my notes. “Ok, well my next

question is, and this is mainly for you, Dr. DP, should I allow students to fail?”

We discussed the pros and cons of either (A) giving students corrective feedback

immediately and setting them back on the right course or (B) allowing students to make incorrect

decisions and following through to the logical conclusion of that decision. For example, say that

a student playing the game reaches anaphase and doesn’t separate the sister chromatids. The end

result of that decision will be one daughter cell with twice the normal amount of genetic

information and the other daughter cell with no genetic information at all, which would not be a

fun time for either of them. This kind of delayed feedback promotes transfer of knowledge and

would provide more possible narrative endings. However, more immediate feedback promotes

the acquisition of knowledge, which is the stage of learning that we intended to target (Schmidt

et al., 1989). Additionally, immediate correction would be more expedient and likely be a better

use of class time, since we intended to use the game as an in-class active learning activity. If the

students were playing on their own time, especially if they could replay it until they won the

game by successfully dividing, giving them various failure endings would likely be more

effective.

“Does it have to be one or the other?” Sue Ellen asked.

“What do you mean?”

“Well, rather than have the students play the entire game after they make a mistake, can

you have a scene where you fast forward to the consequence? Rather than tell them what would

go wrong, show them.”

“Like a flashforward instead of a flashback?” one of the undergrads clarified.

“Hmmm, I might be able to make that work,” I said. “A little bit of the best of both

worlds.”

“Exactly!”

“Next question: What can the pedagogical agent be? A pedagogical agent is a character

that acts like a teacher or guide. They’re an expert in whatever the learning goal is, so what kind

of cell would be an expert in division to be able to fill that role?”

“When I think of fast-growing cells, my first two thoughts are cancer and skin,” Sue

Ellen offered.

“Those were my thoughts also. I don’t want to use a cancerous cell for a few reasons.

They’re not exactly… typical, in terms of how they function and divide. Also, it doesn’t really fit

the vibe of the game,” I explained.

“Good point.”

“A stem cell?” the same undergraduate suggested.

“Yeah I’d thought of that too, but how can I make that make sense narratively? The

player picks what kind of cell they are at the start, so there can’t be any fetal stem cells because

those organs have all developed, and in adults the options for stem cells are pretty limited.

They’re all stuck in their one spot, so they wouldn’t be able to access the parts of the body where

all of the possible cell types are.”

“I think you’re overestimating how much the students know about the human body,” said

Sue Ellen.

“Also a good point,” I said.

“Blood cells come from stem cells, don’t they?” the undergraduate wondered.

“Yeah,” Sue Ellen and I answered at the same time.

“They’re in the bone marrow though,” I continued.

“Are they?” Sue Ellen asked

“...Are they not?” I countered.

“Let’s see,” she said, turning to her computer and opening a new tab. She searched

“hematopoietic stem cells” and read the first entry aloud. “...found in the peripheral blood and

bone marrow. There you go!”

“Ok, now my last problem: you wanted the students to see plant, animal, and fungal cells

in the original version of the game, but the narrative I’ve created takes place in the human body.

Those cells wouldn’t make sense there.”

“No, they wouldn’t,” Sue Ellen agreed. “What I really want more so than for students to

see cells from different domains is for them to just see different sizes and shapes of cells. They

always think a cell is this perfectly spherical, uniform thing, and that’s almost never what cells

look like.”

“So you just want some diversity in shape and function?” I clarified.

“Exactly.”

“So… a muscle cell, that would be long and thin. Cells from intestinal mucosa would be

stratified columnar. What other shapes would be good?”

“Nerve cells are definitely different looking with their dendrites,” Sue Ellen offered.

“But nerves don’t divide,” I said at the same time as one of our undergrads said “I

thought nerves didn’t divide.”

“Hm. That’s right. I didn’t think that thought through.”

We sat in silence for a few seconds, each trying to come up with a cell with a unique

shape we could include in the game. Funnily enough, Sue Ellen’s next door office neighbor at

the time was a neurobiologist, and he was passing her door on his way to his. She called him

over and asked if he had a moment to spare.

“I think I can manage that,” he chuckled. “What do you need?”

“We’re talking about Amanda’s mitosis game for their thesis. We want to show the

students cells of different shapes, and nerve cells have the most different shape we can think of.”

“That they do.”

“But it’s a mitosis game,” I continued. “Nerve cells don’t divide in adults, right?”

“Actually… check out the subventricular zone. Adult neurogenesis does happen there.”

“No way!” some of our undergrads and I said in unison.

“Yes way, promise.”

I left that meeting with a much fuller mental picture of what the game would be, so I first

tried to commit that mental image to paper by sketching out my desired user interface (UI,

shown in Figure 5).

Figure 5. Sketch of UI concept

This was where I started getting truly excited about the game. I had a vision! On paper!

Physically in my hands! I was going for somewhat of a visual novel design, but with more room

for text and cell images. It was time to really start fleshing out the characters and narrative.

First, I made a list of the characters I needed. The player character, obviously. I decided

they would be a cell that had just divided but lost its memories. (How? Doesn’t matter. Cells

don’t have memories anyway. It’s fine, don’t worry about it.) The hematopoietic stem cell,

affectionately dubbed ‘Mat’, had also been decided. In my notes, I marked him as “a sort of

grizzled war vet vibe”, the kind of character that would pick up the player character and go “yeah

you’re screwed, but I got through it and so will you”. I also had some fun in my notes: “[Mat]

guides the player character through the process by offering tips and correction, and a little bit of

humor (of the red variety)”. (Red humor… because he’s a blood cell… blood… the theory of the

four humors? Get it? No? It’s ok, no one in our lab did either. I told you I’m a nerd.)

After the two stars of the show, I listed out the organelles that would need to be

personified to move the narrative. The centrosomes are the movers and shakers of the process, so

write them in for sure. My original idea for them was “a couple of straight-to-the-point, no-

nonsense dudes that come across like factory/assembly line managers”. The chromosomes are

the macromolecules of the hour, so they had to be included, too. From my notes: “Sister

chromatids are copies of each other, so I’m going for a narcissist vibe here where they don’t care

about anything but each other; it’s so over the top and obnoxious that the centromeres can’t

WAIT to pull them apart just so they’ll shut up.” I also wrote in the nuclear envelope, since I

knew students often forgot its role in division. For that, I was aiming for Fezzik in The Princess

Bride: “just a big hulk of a dude that’s very passive, maybe has a good one liner here and there”,

and added: “only going to appear in interphase, then dissolves in prophase, and reforms in

telophase (that last one’s where a one liner is crucial)”. Unfortunately, this idea got lost in the

rush that was trying to get out a script draft and then functional demo. (A note to file away for

possible game improvements?)

Table 12: Summary of brainstorming challenges and decisions

Challenges

Decisions

Building an interesting narrative that

incorporates learning objectives

Collaborative brainstorming with colleagues

and students with necessary content

knowledge

Structuring narrative and game mechanics to

fit within time constraints, needs of students,

and evidence-based GBL practices

Allow “failure” by fast-forwarding to

consequences of player choices within a

single passage, giving corrective feedback,

and returning the player to where the incorrect

decision was made

Transition from script to Twine (due to

including some, but not all, markdown

language and alternative passages in the

script)

Because it took so long to locate and fix the

issues, I would recommend either writing the

script with no regard for how to fit in

markdown language and plan to spend time

(with less frustration) on filling in the

markdown language, or write directly in

Twine based on a detailed outline

Scriptwriting

I started by trying to make an outline based on my flowchart and notes, but quickly

realized that I was getting bogged down in the details and would never get words on paper if I

didn’t just do it. Keeping in mind that I would ultimately end up moving the script into Twine

passages, I included some markdown language in an attempt to ease the transition. To be honest,

I went into a bit of a writer’s fugue state while writing the interphase part of this first draft. I

didn’t take any notes on the process save for this one: “realized while writing that the

centrosomes being too competent takes away a lot of player agency.” In my original draft, the

centrosomes were very leading. They were what I had intended them to be: “a couple of straight-

to-the-point, no-nonsense dudes that come across like factory/assembly line managers”; in other

words, they were good at their job. This didn’t leave a lot of agency for the player, because the

centrosomes knew what they were supposed to do, and when the player would select the wrong

answer, their responses weren’t very constructive. My one note ended with the solution I settled

upon: “also it would be way funnier if they’re more three-stooges style idiots.” I took a little time

to dumb them down in what I had written up to that point and forged ahead with my newly

stooge-ified centrosomes.

To add some interactivity to the interphase section, I added an area where the player is

free to roam the cell. They are given four options: up, down, left, and right, each of which

exposes the player to different structures and characters and allows them to become familiar with

the characters and tools they’ll interact with during mitosis and gives them a fuller picture of the

story setting. This was easy and fun to write, but I didn’t spend much time at this stage thinking

about linking passages and coding variables, which caused a small headache (more on that later).

For mitosis itself, I outlined the phases with brief notes including how I wanted to assess student

knowledge and ideas for actions and dialogue between the characters, then wrote my script

around that. Because at the time Division Quest was intended to be played only in class, I chose

to limit the interactivity, which resulted in more of a kinetic novel vibe that wasn’t my original

intent.

I took my script draft to our next lab meeting, where our undergraduate research students

read it and gave feedback. One of them caught a very big mistake (writing “centromere” where I

meant “centrosome” in several places) which was much easier to fix at this phase (I used find

and replace to locate all instances of the word and correct those that needed correction) than it

would have been in Twine (where I would have had to carefully read every single passage and

correct them one by one), and they both said they would like a little more interactivity. We

brainstormed some ideas together (most notably an interphase minigame where the player fulfills

the function of their chosen cell type), but couldn’t find a solution that (1) would enhance, not

detract from, learning, (2) would be meaningful to the narrative, and (3) would be doable in the

timeframe, both my timeframe for development and the students’ timeframe of playing in class.

The script phase of development was very enjoyable. Creative writing is one of my

favorite hobbies, and it had been a while since I’d made a story for a Twine game. However, I

didn’t allow myself enough time to do it well (I had to drop the flashforward to consequences

idea, for example). That was, in part, due to starting over from scratch after reading the literature.

While I don’t regret doing that, because it resulted in what I feel is a much better product, were

this project not held within the confines of a master’s thesis, I would have preferred to spend

more time with this script, perhaps by making a second draft to building up the characters and

story more. Another factor that restricted the time I spent in this phase was my course load,

teaching schedule, and personal life. In the fall of 2022, when I started reading literature and

made the decision to start over, I was in my first semester of grad school. On top of the

adjustment to a new level of coursework, I was doing my research and meeting with members of

my thesis committee on the Armstrong campus, but taking an in-person lecture with an

associated lab and teaching two in-person lab sections on the Statesboro campus over an hour

away. Making the drive two days a week alone was eating up much of my time and energy, as

was navigating two very different campus cultures. The cherry on the cake was finding out I was

pregnant in October and the physical and emotional toll that took on me. More meetings than I

can count were delayed, rescheduled, or canceled due to how sick I felt that first trimester, and

getting work done even on my own time was a struggle.

By the spring of 2023, many of these issues had improved (I was assigned to teach on the

Armstrong campus, got past the worst of the morning sickness, and had acclimated to graduate

level coursework and scheduling), but I had gotten so behind in my research in the fall that I was

constantly playing catch up. On top of that, I was teaching a majors lab, which no GTAs on the

Armstrong campus had done before, so teaching, even though it took less time than it had in the

fall, still took up a lot of my time and energy.

Table 13: Summary of scriptwriting challenges and decisions

Challenges

Decisions

Characters being too competent diminishes

player agency

Make ‘em stupid!

Typos (especially relating to terminology)

Have others proofread the script before

porting into Twine

Lack of interactivity

For the in-class purposes planned for Division

Quest, a lack of interactivity was considered

acceptable

Playable Demo

The jump from script to first playable demo required two things: completing what syntax

and alternate passages I hadn’t included in the script and bug testing and bug fixing until I

couldn’t find any more bugs. A bug is an unintended event in a program, often due to a mistake

in the code. Bugs I find in my games often are simple typos, missing or incorrectly displayed

images, broken links between passages, and variables not populating as intended and preventing

gameplay from working as expected. This stage was primarily copying and pasting directly from

my script into each Twine passage and tweaking syntax as I went. Without doing anything else,

Twine assigned the default UI design of white text on a black background and blue hyperlinks.

Once the entire script was transferred like this, I did a round of bug testing to make sure

everything was linked correctly and the buttons were displayed the way I wanted them. In

addition to the standard issues (typos, rewording, and fiddling with button layout, all of which

are simple fixes), I found a place where I had apparently been so focused on the narrative that I

neglected to think about how passages would link during the interphase exploration. With the

script pasted in with incomplete syntax, the player could explore the same four areas of the cell

infinitely without ever being given the option to progress the game: an infinite loop, making the

game unplayable beyond that point.

It took a frustrating amount of time to figure out what the problem was, and it’s even

more frustrating in hindsight because of how obvious it seems to me now. Twine displays

passages in a story as boxes labeled with their titles with the links between passages as arrows

between them. Figure 6 shows the section of passages that contained the infinite loop:

Figure 6: Screenshot of a set of Twine passages

I wanted to force the player to explore all four areas of the cell before being allowed to

continue, lest they miss any of the important information in one or more of the areas. That

required the “Exploration Check” passage. At the end of each area, players were given the option

to “take another look around”, which would take them from the current area they were into to the

exploration check. There, they were given buttons to direct them to whatever other areas they

had not already explored. However, once all four areas had been explored, nothing was displayed

at all. I had to program the “Exploration Check” to recognize when all four passages have been

visited and display a new set of text with a button that directs them to the next passage. It was

just a simple oversight due to thinking about the coding a little but not a lot while writing this

script, then being confident that I didn’t need to do any more scrutinizing while copy/pasting.

With this addition, the player would see Mat’s instructions, talk to him, and be given the option

to speak to the centrosomes, which progresses the game. Infinite loop broken!

While including syntax in the script did lead to that headache, it did make it easy to move

everything into Twine. Normally, I would either write directly in Twine or write only the script

in a word processor and do all of the coding in Twine. Doing it this way, I would constantly be

opening the game to test as I go. Because I was (over)confident in my first draft, I wasn’t doing

that testing as often. This doesn’t strike me as an issue unique to programming; I notice this in

many other facets of my life, both work and personal. I choose not to preview a quiz after

changing the settings, and the students can no longer access the quiz at all. I play Among Us with

a group of friends and I’m so sure one of them is the Imposter that I overlook evidence that it’s

someone else. Every once in a while, maybe I just need to be reminded that I’m human and I

make mistakes. and that’s ok.

For the next version, the goal was to improve the visual appeal of the game using the CSS

and add a reference tool to help the player manage their cognitive load.

Table 14: Summary of first playable demo challenges and decisions

Challenges

Decisions

Typos and formatting

Plan time to proofread and troubleshoot

Infinite loop

Be careful and cautious; test the game

frequently while adding passages to check for

errors as they appear

Adding CSS

Manipulating the CSS is the part of making games in Twine that I have the least

experience in and dislike the most (those two things are surely related). Because each version of

Harlowe has slightly different syntax for the CSS, it’s more difficult to find resources for

whichever specific version I’m working with at the time. I can often find a template that’s

exactly or very close to what I’m aiming for, but it’s in a different version, so I either have to

abandon it or try to modify it for the current version (as I did for Original Division Quest). For

this next iteration, I decided not to use the original template, despite all of the work that I spent

updating it to the current version of Harlowe. Thanks to the help of altruistic users on the Twine

Discord server, it looked exactly how I’d originally wanted it: clean; too clean. It felt like a

corporate training module, not a game. However, I did want to keep the template’s “dia” class

that displayed text in a bubble embedded within the larger passage. Luckily, I could cut that

specific class from the old, updated CSS template to incorporate into whatever I used next.

One of my favorite places to find resources like Twine templates is itch.io. Itch.io is an

open marketplace for independent (indie) developers to share and sell their content, so it’s a great

place to find free or low-cost content ranging from individual code elements to complete games.

I searched “twine harlowe template” and found a visual novel template that I liked (funnily

enough right beside the original template in the search results, shown in Figure 7):

Figure 7. Screenshot of two Twine template listings on itch.io: the template used for the new

version of Division Quest (left, found at:

https://sunlabyrinth.itch.io/simple-vn-template-for-twine) and the template used in the original

version (right, found at https://candygiants.itch.io/twine-2-harlowe-css-pack-1)

The character portraits displayed on the side were exactly what I wanted, and the other

elements would be relatively easy to alter, both because of my previous experience doing so and

the programmer’s notes left in the CSS by the template author. Programmer’s notes can be used

to annotate lines of code, most often to explain what the code does to other programmers or to

leave notes for yourself to reference later. And it was in the current version of Harlowe! Perfect.

I did say relatively easy. It would have taken me much, much longer to build anything

like that from scratch, but that didn’t make the modifications a cake walk. Because both my

index link and the character portraits use the sidebar, it made the sidebar look cluttered under the

best conditions, and when there were no portraits being displayed, the space was still being

reserved for them, so the index button floated awkwardly in the middle of the sidebar. It was

functional but not ideal. After exhausting all of my ideas for troubleshooting, I sent it to Rob,

who did some tweaking and sent me a fix that moves the index link into the header and moves

the header out of the story box so it’s displayed separately from where the story text is displayed.

Odd, but clever, and most importantly it worked.

After all of the work to massage the CSS into the form I wanted, I didn’t have time left

before the summer section would be playing it to actually draw the character portraits… so,

while the necessary elements for displaying the portraits were there, what the students actually

saw was blank space. I did, however, have time to create the game index that would serve as a

central knowledge organization tool. To accomplish this, I created a passage linked to the

header-turned-footer that students could click at any time to see the information they had

gathered. However, because I hadn’t yet taught the game how to keep track of that information,

clicking the link took the player to a blank page. I listed everything I wanted the player to be able

to access in the index, then mapped out where they would find that information in the game.

Finally, I added variables to each passage where that information could be found.

Variables are the bread and butter of programming. A variable is a saved value; that value

can be changed under certain, defined conditions and can also be used to instruct a program to

execute specific functions. They can come in a variety of forms. In Twine, variable types

include: boolean variables (where the only possible values are true or false), integer variables

(where the only possible values are numbers), string variables (where the value is a set, or string,

of characters), and arrays (where instead of a single value, the variable contains a list of values).

To make the index, I could use an array, with each piece of information stored in the array’s list,

or I could use many boolean variables, with one for each individual piece of information. The

array would be cleaner on the back end; there would be one variable to keep tabs on, rather than

however many I would need for each index entry. However, boolean variables are simple, and I

have more experience using them, so I went with that. In each passage where a piece of

information was collected, I defined the variable, named it after the piece of information it

represented, and set it to true, which let the program know that the student had collected the

information and it should be displayed in the index. By default, the value is set to false, meaning

the student hasn’t been in the passage where that information is gained and the variable set to

true, so the index entry for it should display “Undiscovered”, but once the student had discovered

it, the variable would be true, and the program would display the entry.

And that’s that! Before sending this off to the students, I sent it to several friends so I had

fresh eyes looking for bugs and typos. One dear, dear friend of mine finds particular joy and

pride in finding the strangest, most easily overlooked bugs in my games. He wanted to know

what would happen if, while in the index, he clicked the link to the index. Surprise, it took him to

the index! He pressed it a few more times, hoping to cause the program to crash or his computer

to self-destruct. When neither happened, he sent me a disappointed message, then continued the

game… or he tried to, at least. When he clicked “Return to last page”, which should have

returned him to whatever passage in the narrative he had left off on, he was pleasantly surprised

to find himself still in the index. I was less pleased with this news.

The “Return to last page” button worked by referencing an array, called (history:), that

contained a list of every passage the player had entered, in the order they entered it. The last

value in the array should be the passage that the player left to enter the index, so the button told

the program to take the player to “(history:)’s last”, but because my friend spammed the index

link while still in the index, the last few entries in his (history:) array looked something like this:

“Index”, “Index”, “Index”. etc.. The last entry in his array was the index, so it took him to the

index. He had completely lost access to whatever part of the story he had been in before entering

the index. That was less than ideal, and I had no idea how to fix it. I tried adding a variable that

tracked whether or not the current passage was the index that I could use to hide the index button

if the player is already in the index, but the player could still access the credits or controls

tutorial page and have the same issue. I reached out to Rob with this dilemma, and fortunately he

had a solution: another variable! Rather than depend on the students not to click the index link

while in the index, I could make a string variable that could hold the title of the current passage

and update when the player entered each new passage. That way, when the player moves to the

index, that string variable will remain the same. Then, I could make the return button reference

that variable rather than (history:)’s last. Problem solved! I just had to paste the new variable into

every single passage. Tedious, but not difficult. I sent the updated version back out to my (very

thorough) playtesters, and there were no more issues.

With a functional and more visually appealing version of the game put together, the game

was ready for its maiden voyage in the classroom. Sue Ellen was teaching one online summer

section of the course, so she embedded the game and a feedback survey in her course materials.

In the survey, I asked questions about the game and the students’ experience with it. I wanted to

know: did the students enjoy the game?; did they feel like it helped them learn (regardless of if it

actually did or not)?; what did they think was good about the game?; and how did they think the

game could be improved? This feedback would inform my next development decisions and give

me time to make alterations over the summer for a second round of student feedback at the

beginning of the fall semester. Student feedback frequencies and example quotes are shown in

Tables 12 through 15. No additional feedback was provided for this version of the game for the

questions “Was the game ever confusing or hard to follow? If so, what part(s)?”, “Did you notice

anything wrong with the game (such as a typo, an error message, not being able to progress

further)”, or “Do you have any additional feedback that was not covered by a previous

question?”

Table 15: Student Responses to “What feature(s), if any, did you find most fun or enjoyable?”

Response

Example Quotes

Narrative

“The fun storyline”

“...[L]ike a story. It was cool.”

Knowledge

Testing

“... [i]t was like playing a trivia game. My favorite game is trivia”

First-person

“Acting as i was the cell”

Casual tone

“It felt like I was reading a comic or watching a cartoon”

“I enjoyed the connections to regular things like hooks and kinetochores”

“I liked that is was like a book monolog and read it like they were actual

people”

Humor

“The remarks were funny”

Table 16: Student Responses to “What feature(s), if any, did you find most helpful for your

learning?”

Response

Example Quotes

Hints and scaffolding

“...edging toward the next step”

“...[i]f you didn’t click the right answer it would give you an

explanation to why it wasn’t right”

“When I was wrong it did not make me feel silly about being wrong”

Casual tone/interactive

conversation

“Making each part of cell division a role and job”

Multiple choices

“Being able to give different answers was helpful”

No failure

“I like how you can keep trying until you get it right”

Index

“Being able to look through the index throughout”

Table 17: Student Responses to: “What parts of the game, if any, made the game less fun and

enjoyable and/or were not helpful for your learning?”

Response

Example Quotes

Excessive length

“Was somewhat long, attention span can get lost”

Lacking visuals

“The game did lack in colors and visual appealing stuff”

“I didn’t like the layout of the game”

None

“I think this has been my favorite way to retain information”

Table 18: Student Responses to: “Would you suggest any changes that would make the game

more enjoyable for you or better help your learning?”

Response

Example Quotes

“Give a little bit more background to the story”

Shorter

“Maybe shorter?”

Add images

“Maybe add images so you also get a visual of the steps”

“Images would help”

“having the visuals of what is being explained shown and maybe have what

was written be the dialogue”

The major criticisms I received for this version of the game were: it needed images,

because the text alone wasn’t very visually engaging, and the amount of text in each passage

wasn’t very balanced. Some passages would have two or three lines of dialogue, while others

had a long paragraph or more. However, they enjoyed the story, thought the characters were

funny, and many felt that the game clarified the process of mitosis for them. One student had

strong criticisms and did not feel like they learned from the game; they felt that it was too linear

and didn’t require any critical thinking. Because almost all of the other students' feedback was

mainly positive, I interpreted this as perhaps the game is not as beneficial for students who

already grasp the fundamentals and are ready for higher levels, such as application, which made

sense because the game was designed to target the lower levels of Bloom’s taxonomy

(remembering and understanding).

It was encouraging that many of the suggestions these students made were things that I

already had on my roadmap. I felt like I was on the right track and was confident in my

decisions. The feedback about the amount of text was less encouraging, since it is a text-based

game, but it was good to know because the game needed to be as accessible as possible. If

students skipped or skimmed sections because they were too long, that wouldn’t do anyone any

good. For the next version, I planned to trim down and separate these larger chunks of text and

add the cell images, character portraits, and illustrations.

Table 19: Summary of CSS addition challenges and decisions

Challenges

Decisions

Out of date template

Find a new template and ensure it’s in the

correct version

Links in sidebar look cluttered with character

portraits

Move links out of sidebar but still separate

from passage (with help from another Twine

user)

Infinite loop if players use resource links

while already using one of the resources

Change method for how “Return to passage”

link works from referencing the previous

passage to storing the last story passage

within a variable

Complicating Factors

It’s important to note that between this version and the last, I had my baby. Adjusting to

life as a parent and surviving the newborn stage became my highest priority, hence why the goals

for this version were minimal. On top of that, Sue Ellen’s teaching schedule had changed, and

she was no longer teaching the course I’d designed the game for. Because of that, I was also

reaching out to and recruiting other instructors on both campuses to use Division Quest in their

sections and help us collect data for the learning outcomes research questions. This also threw a

wrench in the design of the game itself. Decisions that I’d made earlier in development under the

assumption that the game would be used the way Sue Ellen planned to use it didn’t all make

sense, but it was too late to redo anything.

Adding Images and Mobile-Friendliness

The most important images to add were the real cell images. They were the main point of

why I wanted to make the game, after all. The issue was how to implement them. I could have

simply dropped them into the passage and put buttons below each for students to click on, but

that could easily get messy and add to extraneous load. Ideally the students would be able to

click directly on the image and that would take them to the next passage where they receive

feedback on their answer. Luckily, Rob had the same thought and, while helping me format the

header-turned-footer for the index link, sent me some of his demo code that added a hover macro

to the CSS. There, he tells a story about locating a bad smell, and the player clicks on images of

various animals to identify the one that stinks. It was exactly what I wanted to do, and all I had to

do was replace the (button:) macros with the new (hover:) macro and a Dropbox link and alt text

(used for visual impairment accessibility) to the images!

With the cell images embedded, it was a simple task to play through the game myself to

locate the text-heavy passages and edit those down and/or split them into smaller sections. Along

the way, I also did some line editing where I found confusing phrasing or typos, and that was

that! This version of the game was complete with very little headache compared to earlier

versions, and I was done with plenty of time to spare. I sent it to Sue Ellen, who wanted to use it

as a review exercise in her genetics class and collect more student feedback. She responded that

she loved it and was very excited to start using it, but mentioned that some students would likely

be trying to play it on their phones, since she wanted to use it on the first day of class and not

everyone brings laptops to their lectures. She’d asked about the game’s mobile-friendliness while

we were discussing plans not long after my daughter’s birth, but I strongly encouraged her to try

to have them play it on a computer, because making a Twine game mobile-friendly involves

heavily manipulating the CSS, especially the way images are displayed, and I didn’t feel that I

had the skill or the time to do that.

Come fall, when she couldn’t think of a way to make that happen (the course was being

taught across campus, so she couldn’t borrow one of the department’s laptop carts, and we

couldn’t rely on students to check their email or Folio before the first day of class), she suggested

using it in her online section just like we’d done over the summer rather than genetics.

Unfortunately I’m obstinate and really wanted feedback from students who (should) know the

concepts in the game already to see if they might have insight on the phrasing or format, so I

took it upon myself to manhandle the game into a… mobile-functional game, if not fully mobile-

friendly. This mostly last minute decision is, in my opinion, the worst decision I made in the

entire development process. Plus, Sue Ellen forgot to use the QR code with the game link that I

had sent her, so the students didn’t even play it in class. I still got the feedback (shown in Tables

17 through 21); she assigned it as homework… where they could have played on a computer…

Only one comment each was provided for this version of the game for the questions “Did you

notice anything wrong with the game (such as a typo, an error message, not being able to

progress further)” (Response: “‘Gunna’ bothered me”) and “Do you have any additional

feedback that was not covered by a previous question?” (Response: “It’s a good idea, but I think

you can add some more content and focus more on the age group of your college level

audience”).

Table 20: Student responses to: “What feature(s), if any, did you find most fun or enjoyable?”

Response

Example Quotes

Narrative

“I enjoyed the storyline aspect”

“Stories keep the reader more engaged”

Testing

Knowledge

“I liked the parts where you had to make a decision about the stage of cell

division”

Casual tone/

interactive

conversations

“Different and various options”

“I loved how it was like a skit talking back to you”

“Clicking around to see the different selections”

Humor

“I enjoyed the humor”

“I liked that it was playful”

Table 21: Student responses to: “What feature(s), if any, did you find most helpful for your

learning?”

Response

Example Responses

Hints

“The way it refines the answers if you choose one that was incorrect”

“if I got a question wrong it redirected me to the next steps”

Characters

“Having real names in the game to make it more visual”

Casual tone/

interactive

conversations

“It was like a story that I had to ‘write’ and it was fun so it was easier to

follow”

Simple

“The game broke down the subject and made it easy to understand”

No failure

“If I chose the wrong answer, it would tell me to try again”

“Multiple chances”

“It did not matter if you got them wrong it helped you and did not punish you”

Index

“The description of each stage and the ability to refer to the index at the end”

Table 22: Student responses to: “What parts of the game, if any, made the game less fun and

enjoyable and/or were not helpful for your learning?”

Response

Example quotes

None

“There was not a part of the game that felt unenjoyable. It was very easy to

understand”

Length

“It was a lot of words, sometimes it went pass my head”

Lacking

visuals

“No pictures”

“I think it would be helpful to have a visual animation”

“The colors, It kind of washed the screen out”

Lacking

orientation

“It was kind of confusing at the beginning until I figured out what I was

supposed to be doing”

Casual tone

“In some ways, the informal, child-like language made the content more

accessible, but I did feel it was beneath the college level”

Options

“Deciding which option to click on”

Loss of

novelty

“Novelty wear off quickly and it begins to drag out”

Table 23: Student responses to: “Would you suggest any changes that would make the game

more enjoyable for you or better help your learning?”

Response

Example Quotes

Add images

“maybe a picture of the current phase to better have a visual as it goes along”

“include images to grab people’s attention and to remember better”

Add clarification

“More clear directions at the beginning of the game”

Achievements

“More ‘flag pole’ victories. Where you earn something (even just a point or a

star or something)

Table 24: Student responses to: “Was the game ever confusing or hard to follow? If so, what

part(s)?”

Response

Example Quotes

Lots of text

“the long paragraphs”

“It was a lot to read so sometimes it got confusing”

Lack of orientation

“At first until I got used to it”

Following dialogue

“The part where the main person was talking to all the different stages”

Some of the feedback in this version was surprising to me. Much of it reinforced what I

had received in the previous version: students enjoyed the story and characters, especially the

humor and casual tone; the immediate feedback was appreciated; and they enjoyed their

gameplay experience. However, even after trimming down as much text as I felt that I could and

breaking up what I couldn’t trim into smaller pieces, some students still felt that it was a lot of

reading. At this point I assumed that no matter how little reading there were, some students

would still find it too much, and I contented myself with comparing the amount of reading in the

game to that in a textbook, with the game being much shorter, more engaging, and easier to

follow, at least in my opinion.

Additionally, a couple of students mentioned the goal or instructions being confusing,

which none in the previous version had said. With nothing to go on save for their written

responses, I didn’t know what to make of the students who felt lost in the early portions of the

game. Did they read the controls and instructions (which are optional, as I considered the game

to be pretty straightforward, and from the feedback it seems most students agreed with me)?

Were they clicking options without taking time to fully comprehend what they were reading?

Could it be explained by low reading comprehension due to poor K-12 education and/or many

years since leaving K-12?

With only two students reporting this confusion and no way I could see to further reduce

the amount of reading, I decided not to make any alterations based on this feedback. For the final

version, the last major addition was adding illustrations and character portraits, and then I could

do a final sweep to polish it all up.

Table 25: Summary of adding images and mobile-friendliness challenges and decisions

Challenges

Decisions

How to embed cell images

Use hover macro from shared demo code

Students may not have access to a laptop to

play the game in class

Make concessions in design to accommodate

mobile-friendliness

Game is more mobile-functional than mobile-

friendly, and interesting features had to be

abandoned to make it happen

Made concessions out of necessity; for future

work, either plan mobile friendliness from the

beginning or ensure students have access to

appropriate devices

Conflicting student feedback

Examine the areas of feedback and compare

the conflicting sides to determine where, if

any, changes are needed

Adding Illustrations and Finishing Touches

To add illustrations, I needed to have illustrations. I wanted to have character portraits for

Mat, the centrosomes, the sister chromatids, the sisters being separated in anaphase, and scene

illustrations of the four exploration areas early in the game. I already had Mat designed from my

prototype sketch, so I started with making a higher quality version using Infinite Painter on my

Samsung Tab S8 (shown in Figure 8).

Figure 8: Character portrait of Mat, the hematopoietic stem cell pedagogical agent

Next, I did the centrosomes. I wanted them to have a shape reminiscent of actual

centrosomes, but simplified and cartoonized. I made a couple of scribbles with different shapes

and expressions on their faces and settled on a Veggie Tale-esque, childish face (shown in Figure

9).

Figure 9: Character portrait of a centrosome

For the sister chromatids, I needed to find a way to visually depict the fact that they’re a

single duplicated chromosome. I went through even more scribbled sketches of these two, trying

different shapes, poses, and expressions. I ended up with two complete illustrations. One was

their standard portrait, and the other depicted them being pulled apart in anaphase (shown in

Figure 5). The sisters’ arms around each other visually represent the centromere and the hooks

on their backs represent the kinetochore, which are two structures that students commonly

overlook or forget about. Incorporating them into the design was meant to combat that.

Figure 10: Two character portraits of the sister chromatids: their standard portrait, showing them

as a single duplicated chromosome with their arms wrapped tightly around each other (left); and

their anaphase portrait, showing them panicking as they are pulled apart by their kinetochores

(right).

Lastly, I needed to embed those images in the game and, because I was stubborn, made

that mobile friendly. That. Was. A. Nightmare. I tried adjusting the passage width to make room

for the portraits in the sidebar, as I’d originally intended. The passage was pushed entirely off the

screen on mobile. Restricting the image dimensions to a smaller size made most, but not all, of

the passage visible, but the image was so small you could barely see it. In the interest of time and

my sanity, I tried putting the portraits directly in the passage, using the “dia” class as a sort of

speech bubble. Aside from Mat’s portrait being cut off at the bottom (because it would have had

a frame in the sidebar), it looked great on desktop, but when I tried it on my phone, everything

was horizontally condensed, making the speech bubble unreadable. Again, in the interest of time

and my sanity, I did away with the speech bubbles and just dropped the images into the passage.

The only problem that left was the images being off-center, even though I formatted them to be

centered, which I couldn’t be bothered to fix at that point. Also, because the character portraits

took more time than I had planned and embedding them in the game took even longer, I ran out

of time to complete four full scenery illustrations and had to drop them.

Table 26: Summary of adding illustrations and final touches challenges and decisions

Challenges

Decisions

Not having time to make illustrations

Don’t include illustrations :(

Summary and Final Reflection

In addition to modeling the ideas, goals, challenges, and solutions involved in developing

this kind of GBL tools, our autoethnography reveals a set of overarching challenges to

developing and implementing a GBL tool. First, it highlights the importance of knowing where

and how the game will be used while making early development decisions. Second, it exposes

the value of building games collaboratively with peers as well as incorporating feedback from

students. Third, it places an emphasis on starting small, gathering peer and student feedback

frequently, and incorporating feedback into stepwise changes in an ongoing, iterative process.

As the developer, it is especially enlightening to look back on each stage of development

and see how my vision for Division Quest changed over time. While there are features that I

would like to revisit in future versions, such as the cell scenery illustrations, other features that I

was initially excited to include but could not due to time constraints, such as the flashforward to

failure conditions, may have been problematic. Students reporting the game’s casual approach to

corrective feedback to be a contributor to their learning and enjoyment. If they instead were sent

to the natural consequences of their incorrect decisions, that may have had a negative impact on

affect.

CHAPTER 4

DISCUSSION

Summary of Quantitative Results

By using this mixed methods approach, we are able to better inform our interpretation of

statistical results by drawing from the autoethnography results to contextualize and therefore

broaden our understanding of how this kind of GBL tool can overcome some difficulties of

mitosis learning. In summary, Division Quest was demonstrated to improve students’ ability to

recognize a duplicated chromosome and identify its nature as a single chromosome. However,

students who played Division Quest struggled more to visually identify the phase of a cell from a

micrograph compared to students who completed the instructor activities. On all other

assessment criteria (remembering the sequence of the cell cycle; understanding that duplicated

chromosomes become two individual chromosomes upon being separated and when that

happens; and recognizing cell phase from a micrograph, understanding what is happening in that

phase, and using that information to predict what will happen next), students who played

Division Quest learned comparably to students who completed the instructor activities.

Interpretation of Quantitative Results Based on Autoethnography

In feedback during development, many students reported the casual tone of the narrative

and character dialogue as factors that contributed to their learning. Therefore, it is possible that

the casual presentation, discussion of, and interactions with the sister chromatids in Division

Quest contributed to the increased learning surrounding visually identifying and understanding

the nature of duplicated chromosomes. In particular, the illustrations of the sister chromatids,

which depict the sister locked in a tight embrace, and the casual use and defining of terminology

such as centromere, kinetochore, chromatid, and chromosome within the narrative may have

helped the students better integrate this knowledge into their mental models.

However, it was unexpected that students who played Division Quest were less able to

visually identify cell phase from a micrograph, as this was one of the core learning objectives the

game was designed to target. While it is possible that this result can be attributed to extraneous

load from any number of game features (e.g. excessive narrative immersion or lack of

insufficient pretraining of navigating game mechanics), closer inspection of instructor materials,

game content, and assessment materials reveal that the cell images provided in the instructor

materials more closely match those in the assessment compared to those in Division Quest.

Knowing this, it is likely that the content of Division Quest, which was designed to target the

acquisition stage of learning, did not adequately promote transfer of knowledge beyond

recognizing a specific type of cell.

Next Steps for Division Quest

To enhance Division Quest to further promote learning, we have identified several

possible improvements to game features based on our mixed methods analysis. First, the addition

of more illustrations, and possibly some animations, could help students integrate mitosis

concepts into their mental models by offering written or visual information that is relatable to

their lived experiences. To explore whether or not the observed difference in learning

surrounding duplicated chromosomes is due to the character illustrations, a version of the game

without character illustrations and a version of the game with more illustrations and/or

animations can be created and compared through a value-added approach.

To correct the misalignment of cell images between game content and assessment

materials, there are two possible approaches: either swap out cell images so game content and

assessment materials use similar cell images, or diversify cell images within the game so that, in

theory, students will be able to transfer their knowledge of mitotic phases regardless of what type

of cell they are shown. While the former would keep the game focused on acquisition of

knowledge, the latter could result in higher learning gains; however, including a variety of cell

images could also contribute to cognitive load and detract from learning. Value-added studies

exploring game versions where only cell images differ would be needed.

Limitations

The present study’s small sample size resulted in low statistical power and therefore

inhibits our ability to draw generalizable conclusions based on our findings. Likewise, even

though MANCOVAs and ANCOVAs are robust, not all assumptions were met, and the lack of

demographic data for our control group prevented us from performing analyses to determine if

results vary between students of different races, ethnicities, and genders. Changes made in the

timing of survey delivery and the inclusion of instructor materials alongside Division Quest for

some participants also introduce complicating factors. Further studies using a larger sample size

and tighter delivery of pre/post tests are required to determine the validity of our findings.

79 
 
 
CHAPTER 5 
CONCLUSION 
  Our findings align with those in previous studies on GBL design and features. The 
misalignment between the game content and course learning objectives across two introductory 
biology courses reinforces the observation made by Homer et al. (2020) that GBL tools should 
target specific learning objectives and skills, and therefore may not be suitable for multiple 
contexts, such as in- versus out-out-class. Additionally, our results on dissimilar cell images 
between the game and assessment materials reinforces the argument made by (Pawar et al., 
2020) that careful adherence to learning objectives and assessment criteria is vital while 
developing effective GBL tools. Lastly, our findings on the use of character illustrations and 
casual tone in introduction and use of terminology support findings that multimodal information 
delivery (in this case, written and illustrated) promotes learning by utilizing multiple cognitive 
pathways (Mayer, 2020).  
  Similarly, our findings support the use of Twine GBL tools for mitosis learning. Further 
study is required to improve the game, perhaps through versions of the game designed for 
specific learning contexts, such as an in-class game with enhanced outer loop mechanics, which 
are game mechanics that dynamically alter learning content to provide personalized scaffolding 
for students at all stages of learning (Shaker et al., 2016; VanLehn, 2006), or an out-of-class 
game that includes multiple “failure” and “success” conditions or randomized encounters to 
allow for iterative learning sessions through replayability, thereby promoting transfer of 
knowledge and deeper understanding (Schmidt et al., 1989). Additionally, our findings provide a 
foundation for similar Twine GBL tools for mitosis learning targeting other learning objectives 
or using different GBL features. 

For the current game, there are a few ways we can approach improving upon and

exploring the use of Division Quest, but first, we must decide what niche we want the game to

fill, namely who wants to use it and in what context(s) and make changes to suit that. One

instructor expressed interest in using the game as an out-of-class active learning activity to

maximize in-class time in a different way. To best serve this purpose, we could develop the

branching narrative structure that was planned for Division Quest but was abandoned due to lack

of time for development. Regardless of use in- or out-of-class, we believe the game would

benefit by adding outer loop mechanics and stealth assessments (analysis of player choices to

assess learning indirectly) that would allow the game to adjust dynamically and personalize

learning content and feedback to the learning stage of each student, as well as enable data

collection of stealth assessments to inform the instructor of the students’ learning (which may or

may not be assigned a grade). For targeted use as a pre-lab activity, we would incorporate lab

skills into the game, such as the parts of and operation of a compound microscope, and focus

more heavily on diversifying cell images and offering repeated opportunities to practice and

demonstrate proficiency in visually identification of mitotic phase. Because Twine is open-

source and adaptable, we are not necessarily restricted to a single path. Copies of the game in its

current state could be made and altered (by one developer or multiple) in the same stepwise

fashion demonstrated in this study to build toward one or more of these possible uses.

For others looking to take the GBL development methods demonstrated here into

practice, we have three recommendations based on our findings. First, attempt to connect with

other instructors with the same interest in generating individualized GBL tools to work with

collaboratively (either developing together or sharing feedback with each other). Second,

connect with the Twine community and utilize the wealth of knowledge and resources in the

Twine FOSS community. Finally, start small; experiment with Twine’s capabilities and using it

as a medium for your teaching preferences and course structure. Gather feedback from peers and

students regularly to inform development decisions. Make small changes as you are able and see

the need until you are satisfied with the result, and share your work with others to use and

modify. Perhaps eventually, we can collectively build a Twine GBL FOSS community.

82 
 
 
REFERENCES 
Aldahmash, A. H., & Alshaya, F. S. (2012). Secondary School Students’ Alternative 
Conceptions about Genetics. Electronic Journal of Science Education, 16(1). 
Anguera, J. A., Boccanfuso, J., Rintoul, J. L., Al-Hashimi, O., Faraji, F., Janowich, J., Kong, E., 
Larraburo, Y., Rolle, C., Johnston, E., & Gazzaley, A. (2013). Video game training 
enhances cognitive control in older adults. Nature, 501(7465), 97–101. 
https://doi.org/10.1038/nature12486 
Benjamin S. Bloom. (1984). Taxonomy of Educational Objectives, Handbook 1: Cognitive 
Domain (2nd ed.). Longman Publishing Group. 
Bochner, A. P., & Ellis, C. (2022). Why Autoethnography? Social Work and Social Sciences 
Review, 23(2), 8–18. https://doi.org/10.1921/swssr.v23i2.2027 
Bryant, F. B., Kastrup, H., Udo, M., Hislop, N., Shefner, R., & Mallow, J. (2013). Science 
Anxiety, Science Attitudes, and Constructivism: A Binational Study. Journal of Science 
Education and Technology, 22(4), 432–448. https://doi.org/10.1007/s10956-012-9404-x 
Chattopadhyay, A. (2005). Understanding of Genetic Information in Higher Secondary Students 
in Northeast India and the Implications for Genetics Education. Cell Biology Education, 
4(1), 97–104. https://doi.org/10.1187/cbe.04-06-0042 
Cleveland, L. M., Olimpo, J. T., & DeChenne-Peters, S. E. (2017). Investigating the Relationship 
between Instructors’ Use of Active-Learning Strategies and Students’ Conceptual 
Understanding and Affective Changes in Introductory Biology: A Comparison of Two 
Active-Learning Environments. CBE—Life Sciences Education, 16(2), ar19. 
https://doi.org/10.1187/cbe.16-06-0181 
Demirci, N. (2003). Dealing with misconceptions about force and motion concepts in physics: A 

83 
 
 
study of using web-based physics program. Hacettepe Üniversitesi Eğitim Fakültesi 
Dergisi, 24, 40–47. 
Dikmenli, M. (2010). Misconceptions of cell division held by student teachers in biology: A 
drawing analysis. Scientific Research and Essay, 5(2), 235–247. 
England, B. J., Brigati, J. R., Schussler, E. E., & Chen, M. M. (2019). Student Anxiety and 
Perception of Difficulty Impact Performance and Persistence in Introductory Biology 
Courses. CBE—Life Sciences Education, 18(2), ar21. https://doi.org/10.1187/cbe.17-12-
0284 
Freeman, S., Eddy, S. L., McDonough, M., Smith, M. K., Okoroafor, N., Jordt, H., & 
Wenderoth, M. P. (2014). Active learning increases student performance in science, 
engineering, and mathematics. Proceedings of the National Academy of Sciences, 
111(23), 8410–8415. https://doi.org/10.1073/pnas.1319030111 
Georgia Southern University Fact Book 2022-2023. (n.d.). Retrieved April 19, 2024, from 
https://drive.google.com/file/d/13IebXaeLplcIK6hvG2ZJV71U_YzRTy44/view?usp=em
bed_facebook 
Homer, B. D., Plass, J. L., Raffaele, C., Ober, T. M., & Ali, A. (2018). Improving high school 
students’ executive functions through digital game play. Computers & Education, 117, 
50–58. https://doi.org/10.1016/j.compedu.2017.09.011 
Homer, B. D., Raffaele, C., & Henderson, H. (2020). Games as Playful Learning: Implications of 
Development Theory for Game-Based Learning. In Handbook of Game-Based Learning. 
The MIT Press. 
Johnson, C. I., & Mayer, R. E. (2010). Applying the self-explanation principle to multimedia 
learning in a computer-based game-like environment. Online Interactivity: Role of 

84 
 
 
Technology in Behavior Change, 26(6), 1246–1252. 
https://doi.org/10.1016/j.chb.2010.03.025 
Jones, S. M., Katyal, P., Xie, X., Nicolas, M. P., Leung, E. M., Noland, D. M., & Montclare, J. 
K. (2019). A ‘KAHOOT!’ Approach: The Effectiveness of Game-Based Learning for an 
Advanced Placement Biology Class. Simulation & Gaming, 50(6), 832–847. 
https://doi.org/10.1177/1046878119882048 
Keleş, U. (2022). Writing a “Good” Autoethnography in Educational Research: A Modest 
Proposal. The Qualitative Report. https://doi.org/10.46743/2160-3715/2022.5662 
Klymkowsky, M. W. (2010). Thinking about the Conceptual Foundations of the Biological 
Sciences. CBE—Life Sciences Education, 9(4), 405–407. https://doi.org/10.1187/cbe.10-
04-0061 
Lester, D., Skulmoski, G. J., Fisher, D. P., Mehrotra, V., Lim, I., Lang, A., & Keogh, J. W. L. 
(2023). Drivers and barriers to the utilisation of gamification and game‐based learning in 
universities: A systematic review of educators’ perspectives. British Journal of 
Educational Technology, 54(6), 1748–1770. https://doi.org/10.1111/bjet.13311 
Lewis, J., Leach, J., & Wood-Robinson, C. (2000). Chromosomes: The missing link—Young 
people’s understanding of mitosis, meiosis, and fertilisation. Journal of Biological 
Education, 34(4), 189–199. 
Lewis, J., & Wood-Robinson, C. (2000). Genes, Chromosomes, Cell Division and Inheritance—
Do Students See Any Relationship? International Journal of Science Education, 22, 177–
195. http://dx.doi.org/10.1080/09500690210126793 
Litten, K., & Stewart, M. P. (2023). Implementing a choose your own adventure activity to 
improve insulin decision making. Currents in Pharmacy Teaching and Learning, 15(2), 

85 
 
 
149–154. https://doi.org/10.1016/j.cptl.2023.02.020 
Mallow, J. V., & Greenburg, S. L. (1983). Science anxiety and science learning. The Physics 
Teacher, 21(2), 95–99. https://doi.org/10.1119/1.2341214 
Mayer, R. E. (2009). Multimedia learning (2nd ed.). Cambridge University Press. 
Mayer, R. E. (2020). Cognitive foundations of game-based learning. In Handbook of game-based 
learning. (pp. 83–110). The MIT Press. 
Mayer, R. E., & Estrella, G. (2014). Benefits of emotional design in multimedia instruction. 
Learning and Instruction, 33, 12–18. https://doi.org/10.1016/j.learninstruc.2014.02.004 
Mayer, R. E., & Johnson, C. I. (2010). Adding instructional features that promote learning in a 
game-like environment. Journal of Educational Computing Research, 42(3), 241–265. 
https://doi.org/10.2190/EC.42.3.a 
McDonald, K., & Gomes, J. (2013). Evaluating Student Preparedness and Conceptual Change in 
Introductory Biology Students Studying Gene Expression. Journal of Transformative 
Leadership & Policy Studies, 3(1), 21–34. https://doi.org/10.36851/jtlps.v3i1.460 
Morningstar-Kywi, N., & Kim, R. E. (2021). Using Interactive Fiction to Teach Clinical 
Decision-Making in a PharmD Curriculum. Medical Science Educator, 31(2), 687–695. 
https://doi.org/10.1007/s40670-021-01245-7 
Murphy, K. J., Griffin, L. L., Nolan, G., Haigh, A., Hochstrasser, T., Ciuti, S., & Kane, A. 
(2022). Applied autoethnography: A method for reporting best practice in ecological and 
environmental research. Journal of Applied Ecology, 59(11), 2688–2697. 
https://doi.org/10.1111/1365-2664.14252 
Noor Shaker, Julian Togelius, & Mark J. Nelson. (2016). Procedural Content Generation in 
Games (1st ed.). Springer. 

86 
 
 
Ozcan, T., Yildirim, O., & Ozgur, S. (2012). Determining of the University Freshmen Students’ 
Misconceptions and Alternative Conceptions about Mitosis and Meiosis. Procedia - 
Social and Behavioral Sciences, 46, 3677–3680. 
https://doi.org/10.1016/j.sbspro.2012.06.126 
Parong, J., Mayer, R. E., Fiorella, L., MacNamara, A., Homer, B. D., & Plass, J. L. (2017). 
Learning executive function skills by playing focused video games. Contemporary 
Educational Psychology, 51, 141–151. https://doi.org/10.1016/j.cedpsych.2017.07.002 
Pawar, S., Tam, F., & Plass, J. L. (2020). Emerging Design Facotrs in Game-Based Learning: 
Emotional Design, Musical Score, and Game Mechanics Design. In Handbook of Game-
Based Learning. MIT Press. 
Plass, J. L., Goldman, R., Flanagan, M., & Perlin, K. (2009). Improving Self-Efficacy and Self-
Esteem With an Educational Computer Game. . . Hong Kong. 
Plass, J. L., Homer, B. D., Mayer, R. E., & Kinzer, C. K. (2020). Theoretical foundations of 
game-based and playful learning. In Handbook of game-based learning. (pp. 3–24). The 
MIT Press. 
Plass, J. L., Mayer, R. E., & Homer, B. D. (2020). Handbook of game-based learning. The MIT 
Press. 
Rahma, B., Moncef, Z., Boujemaa, A., Benjelloun, N., Anouar, A., & Lhoussaine, M. (2022). 
UNIVERSITY STUDENTS’ KNOWLEDGE AND MISCONCEPTIONS ABOUT 
CELL STRUCTURE AND FUNCTIONS. European Journal of Education Studies, 
9(10). https://doi.org/10.46827/ejes.v9i10.4494 
Riemeier, T., & Gropengießer, H. (2008). On the Roots of Difficulties in Learning about Cell 
Division: Process‐based analysis of students’ conceptual development in teaching 

87 
 
 
experiments. International Journal of Science Education, 30(7), 923–939. 
https://doi.org/10.1080/09500690701294716 
Salen, K., & Zimmerman, E. (2003). Rules of Play. MIT Press. 
Schmidt, H. G., Wagener, S. L., Smeets, G. A. C. M., Keemink, L. M., & Van Der Molen, H. T. 
(2015). On the Use and Misuse of Lectures in Higher Education. Health Professions 
Education, 1(1), 12–18. https://doi.org/10.1016/j.hpe.2015.11.010 
Schmidt, R. A., Young, D. E., Swinnen, S., & Shapiro, D. C. (1989). Summary knowledge of 
results for skill acquisition: Support for the guidance hypothesis. Journal of Experimental 
Psychology: Learning, Memory, and Cognition, 15(2), 352–359. 
https://doi.org/10.1037/0278-7393.15.2.352 
Shaffer, D. W., Squire, K. R., Halverson, R., & Gee, J. P. (2005). Video Games and the Future of 
Learning. Phi Delta Kappan, 87(2), 105–111. 
https://doi.org/10.1177/003172170508700205 
Stains, M., Harshman, J., Barker, M. K., Chasteen, S. V., Cole, R., DeChenne-Peters, S. E., 
Eagan, M. K., Esson, J. M., Knight, J. K., Laski, F. A., Levis-Fitzgerald, M., Lee, C. J., 
Lo, S. M., McDonnell, L. M., McKay, T. A., Michelotti, N., Musgrove, A., Palmer, M. 
S., Plank, K. M., … Young, A. M. (2018). Anatomy of STEM teaching in North 
American universities. Science, 359(6383), 1468–1470. 
https://doi.org/10.1126/science.aap8892 
Stepans, J. (2003). Targeting Students’ Science Misconceptions: Physical Science Concepts 
Using the Conceptual Change Model. Idea Factory. 
Sweller, J., Ayres, P., & Kalyuga, S. (2011). Cognitive load theory. Springer. 
Sweller, J., Van Merrienboer, J. J. G., & Paas, F. G. W. C. (1998). Cognitive Architecture and 

88 
 
 
Instructional Design. Educational Psychology Review, 10(3), 251–296. 
https://doi.org/10.1023/A:1022193728205 
Szilas, N., & Acosta, M. E. (2011). A Theoretical Background for Educational Video Games: 
Games, Signs, Knowledge. In . Handbook of Research on Improving Learning and 
Motivation through Educational Games: Multidisciplinary Approaches (pp. 215–238). 
IGI Global. https://doi.org/10.4018/978-1-60960-495-0 
Terry, R., & Dusenberry, L. (2018). Serious Interactive Fiction: Constraints, Interfaces, and 
Creative Writing Pedagogy. Journal of Creative Writing Studies, 3. 
Twine / An open-source tool for telling interactive, nonlinear stories. (n.d.). Retrieved April 18, 
2024, from https://twinery.org/ 
VanLehn, K. (2006). The Behavior of Tutoring Systems. Internation Journal of Articial 
Intelligence in Education, 16, 227–265. 
Vygotsky, L. (1962). Thought and language. (pp. xxi, 168). MIT Press. 
https://doi.org/10.1037/11193-000 
Watson, W. R., & Yang, S. (2016). Games in Schools: Teachers’ Perceptions of Barriers to 
Game-based Learning. Journal of Interactive Learning Research, 27(2), 153–170. 
Wirtz, B. (2023, July 28). Postmortem Games Analysis: Learning and Improving. Game 
Designing. https://www.gamedesigning.org/learn/postmortem/ 
   

APPENDIX

Table 1: Five-question pre/post test used to assess mitosis learning

Question

Images (if

applicable)

Learning Objective and Bloom’s level

How many chromosomes

are in the middle cell?

a. 1 c. 3

b. 2 d. 4

–Remembering

–Describe the appearance of the genetic material

during each phase of the cell cycle

Which of the following is true concerning

mitosis and the cell cycle? Mitosis occurs…

a. immediately before S phase

b. immediately after cytokinesis

c. in between G1 and G2

d. after interphase

–Remembering

–List the phases of the cell cycle and describe

the sequence of events that occurs during each

phase

During which phase of mitosis do the

duplicated chromosomes split into single

chromosomes?

a. Prometaphase

b. Anaphase

c. Prophase

d. metaphase

–Understanding

–Identify key processes within the phases of the

cell cycle

–Describe the major events of each phase of

mitosis and cytokinesis in animal cells

–Describe the amount of genetic material that is

present during each phase of the cell cycle

–List the phases of mitosis and describe the

events characteristic of each phase

Which of the following

human cells is in prophase?

a. A d. D

b. B e. E

c. C

–Understanding

–Identify stages of mitosis

–Identify the stage of mitosis when

chromosomes condense

–Recognize the phases of mitosis from diagrams

and micrographs

Which event in mitosis will

occur next?

a. Synthesis of

chromatids

b. Spindle fiber

formation

c. Formation of

telophase nuclei

d. Nuclear envelope

breakdown

–Application

–Identify stages of mitosis

–Recognize the phases of mitosis from diagrams

and micrographs

–Identify key processes within the phases of the

cell cycle

–Describe the major events of each phase of

mitosis and cytokinesis in animal cells

–List the phases of mitosis and describe the

events characteristic of each phase

Table 2

Demographic information of Georgia Southern University, Fall of

2022

Sample Characteristics

Home Campus

Statesboro

65.5%

Armstrong

17.1%

Online

10.2%

Liberty

1.5%

Race

White

58.5%

Black

25.6

Hispanic

7.9%

Multiracial

4.4%

Asian

2.6%

Unknown

0.7%

Continental Indigengous American

0.3%

Pacific Islander

0.1%

Gender

Women

59.5%

Men

40.5%

Total

25,000 students

Table 3

Demographic characteristics for experimental group participants.

Campus

Armstrong

Statesboro

Sample Characteristics

Sample

Characteristics

Race

Black

14.3

White

49.1

White

43.9

Latino

5.4

Latino

2.6

Asian

9.1

Asian

Multiracial

9.1

Multiracial

0.5

Other

1.1

No response

7.3

No response

37.6

Gender

Woman

67.3

Woman

49.7

Man

23.5

Man

12.2

Nonbinary

3.6

Nonbinary

0.5

No response

7.3

No response

37.6

Note: N

=55, N

=189.

Table 4: Testing Assumptions for Comparison of Content Knowledge Between Instructor

Activities and Division Quest

Box’s M

df1

df2

Box’s Text

51.137

1.087

74177.271

0.319

Levene’s Test

4.982

245

0.002*

9.463

245

<0.001*

7.750

245

<0.001*

6.428

245

<0.001

0.174

245

0.912

Total

1.049

245

0.372

Note: * indicates statistical significance and therefore failure to meet test assumption

Table 5

Test statistics for Comparison of Content Knowledge Between Instructor Activities and

Division Quest

EMM

SEM

EMM

SEM

df b/t

df w/i

0.234

0.044

0.363

0.035

5.155

248

0.024

0.622

0.050

0.062

0.040

0.001

248

0.976

0.701

0.049

0.590

0.039

3.078

248

0.081

0.669

0.052

0.468

0.041

9.011

248

0.003

0.520

0.052

0.426

0.041

1.966

248

0.162

Total

0.546

0.027

0.499

0.021

1.807

248

0.180

Note: EMM = estimated marginal mean and SEM = standard error of the mean, A = Instructor

activities, B = Division Quest.

Table 6

Testing Assumptions for Comparison of Content Knowledge Between Between In-Class and

Out-of-class

Box’s M

df1

df2

Box’s Text

15.818

1.030

151316.637

0.420

Levene’s Test

4.321

247

0.039*

19.057

247

<0.001*

25.848

247

<0.001*

1.234

247

0.268

0.299

247

0.585

Total

0.121

0.729

Note: * indicates statistical significance and therefore failure to meet test assumption

Table 7

Test statistics for Comparison of Content Knowledge Between In-Class and Out-of-Class

EMM

SEM

EMM

SEM

df b/t

df w/i

0.315

0.045

0.312

0.034

0.002

248

0.967

0.679

0.050

0.563

0.038

3.272

248

0.072

0.737

0.049

0.528

0.038

10.895

248

0.001

0.541

0.053

0.549

0.040

0.017

248

0.897

0.430

0.052

0.494

0.040

0.889

248

0.347

Total

0.551

0.027

0.494

0.022

2.583

248

0.109

Note: Note: EMM = estimated marginal mean and SEM = standard error of the mean, A = In-

Class, B = Out-of-Class.

Table 8.

Cross tabulation comparing game completion across learning contexts

Self-reported level of game completion

Learning context

Full Completion

Partial

Completion

Never played

Total

At-home pre-lab

assignment

118

In-class active

learning activity

Total

169

Note: Pearson Chi-Square = 50.444, df = 2, p = < .001

Table 9

Testing Assumptions for Comparison of Content Knowledge Between Students Who Fully

Completed Division Quest and Students Who Never Played Division Quest

Box’s M

df1

df2

Box’s Text

11.717

0.715

12068.134

0.772

Levene’s Test

0.012

0.914

0.024

0.878

0.904

0.345

1.038

0.312

4.2

0.044*

Total

0.599

0.442

Note: * indicates statistical significance and therefore failure to meet test assumption